Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052919
A. Ferrante, G. Molfetta
Computer vision and image processing are increasingly being used as investigative tools in various fields of science. Compared to other types of measurement techniques, image processing of high-speed camera shooting allows to obtain semi-quantitative information on dynamic aspects of the phenomenon under investigation. Combustion, both for propulsion and for energy production, is characterized by very fast phenomena that cannot be revealed by the typical measurement techniques used in test rigs for the development of prototypes. These dynamic phenomena, often called flame dynamics or flame instability, heavily affect the performance of modern combustors and their study has become essential, not only for the scientific knowledge of the phenomena, but also for the development of new combustion techniques. Therefore, for more than twenty years, the scientific community has been using fast imaging techniques to reveal phenomena that occur during combustion and this has allowed us to deepen our knowledge of the complex phenomena related to combustion itself. This paper describes the use of fast imaging and image processing techniques for the investigation of flame instability phenomena generated by gas turbine burners in an atmospheric test rig where a full-scale burner is tested. Optical investigations are conducted in the visible region of the electromagnetic spectrum.
{"title":"Innovative tools for investigation on flame dynamics by means of fast imaging","authors":"A. Ferrante, G. Molfetta","doi":"10.1109/IPAS55744.2022.10052919","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052919","url":null,"abstract":"Computer vision and image processing are increasingly being used as investigative tools in various fields of science. Compared to other types of measurement techniques, image processing of high-speed camera shooting allows to obtain semi-quantitative information on dynamic aspects of the phenomenon under investigation. Combustion, both for propulsion and for energy production, is characterized by very fast phenomena that cannot be revealed by the typical measurement techniques used in test rigs for the development of prototypes. These dynamic phenomena, often called flame dynamics or flame instability, heavily affect the performance of modern combustors and their study has become essential, not only for the scientific knowledge of the phenomena, but also for the development of new combustion techniques. Therefore, for more than twenty years, the scientific community has been using fast imaging techniques to reveal phenomena that occur during combustion and this has allowed us to deepen our knowledge of the complex phenomena related to combustion itself. This paper describes the use of fast imaging and image processing techniques for the investigation of flame instability phenomena generated by gas turbine burners in an atmospheric test rig where a full-scale burner is tested. Optical investigations are conducted in the visible region of the electromagnetic spectrum.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114463348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052971
Surajsingh Dookhee
The overwhelming amount of household solid waste generated daily is alarming, and this contributes to the rise in pollution and drastic climate change. In such a context, automated waste classification at the initial stage of disposal can be an effective solution to separate recyclable items. Convolutional Neural Networks based on deep learning are often used for automated waste classification, but however, research works are limited to insufficient categories of waste such as the TrashNet dataset consisting of 2,527 images and 6 categories of waste. This dataset does not include other important categories such as battery, biological, and clothing items to reflect real-life environmental problems. Therefore, in this paper, a larger dataset consisting of 15,515 images and 12 categories of common household solid waste was used to evaluate the performance of DenseNet121, DenseNet169, EfficientNetB0, InceptionV3, MobileNetV2, ResNet50, VGG16, VGG19, and Xception Convolutional Neural Network models. Data augmentation was applied to solve the problem of class imbalance, and findings of my first research showed that the Xception model compiled with Adam optimiser outperformed all other models with an accuracy of 88.77% and an F1-score of 0.89. The performance of the model was improved to 89.57% with an F1-score of 0.90 when compiled with Nadam optimiser. However, further experimentation showed that the model did not generalise well despite reaching an accuracy of 93.42% and an F1-score of 0.93 when trained without data augmentation. This demonstrates the feasibility of the proposed model for real-life environmental problems.
{"title":"Domestic Solid Waste Classification Using Convolutional Neural Networks","authors":"Surajsingh Dookhee","doi":"10.1109/IPAS55744.2022.10052971","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052971","url":null,"abstract":"The overwhelming amount of household solid waste generated daily is alarming, and this contributes to the rise in pollution and drastic climate change. In such a context, automated waste classification at the initial stage of disposal can be an effective solution to separate recyclable items. Convolutional Neural Networks based on deep learning are often used for automated waste classification, but however, research works are limited to insufficient categories of waste such as the TrashNet dataset consisting of 2,527 images and 6 categories of waste. This dataset does not include other important categories such as battery, biological, and clothing items to reflect real-life environmental problems. Therefore, in this paper, a larger dataset consisting of 15,515 images and 12 categories of common household solid waste was used to evaluate the performance of DenseNet121, DenseNet169, EfficientNetB0, InceptionV3, MobileNetV2, ResNet50, VGG16, VGG19, and Xception Convolutional Neural Network models. Data augmentation was applied to solve the problem of class imbalance, and findings of my first research showed that the Xception model compiled with Adam optimiser outperformed all other models with an accuracy of 88.77% and an F1-score of 0.89. The performance of the model was improved to 89.57% with an F1-score of 0.90 when compiled with Nadam optimiser. However, further experimentation showed that the model did not generalise well despite reaching an accuracy of 93.42% and an F1-score of 0.93 when trained without data augmentation. This demonstrates the feasibility of the proposed model for real-life environmental problems.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"138 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114832963","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052868
A. Alsawy, Alan Hicks, Dan Moss, Susan Mckeever
Autonomous delivery-by-drone of packages is an active area of research and commercial development. However, the assessment of safe dropping/ delivery zones has received limited attention. Ensuring that the dropping zone is a safe area for dropping, and continues to stay safe during the dropping process is key to safe delivery. This paper proposes a simple and fast classifier to assess the safety of a designated dropping zone before and during the dropping operation, using a single onboard camera. This classifier is, as far as we can tell, the first to address the problem of safety assessment at the point of delivery-by-drone. Experimental results on recorded drone videos show that the proposed classifier provides both average precision and average recall of 97% in our test scenarios.
{"title":"An Image Processing Based Classifier to Support Safe Dropping for Delivery-by-Drone","authors":"A. Alsawy, Alan Hicks, Dan Moss, Susan Mckeever","doi":"10.1109/IPAS55744.2022.10052868","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052868","url":null,"abstract":"Autonomous delivery-by-drone of packages is an active area of research and commercial development. However, the assessment of safe dropping/ delivery zones has received limited attention. Ensuring that the dropping zone is a safe area for dropping, and continues to stay safe during the dropping process is key to safe delivery. This paper proposes a simple and fast classifier to assess the safety of a designated dropping zone before and during the dropping operation, using a single onboard camera. This classifier is, as far as we can tell, the first to address the problem of safety assessment at the point of delivery-by-drone. Experimental results on recorded drone videos show that the proposed classifier provides both average precision and average recall of 97% in our test scenarios.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123907671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10053039
Débora N.P. Oliveira, Marcos R. A. Morais, Antonio M. N. Lima
This paper deals with the design of an indoor optical tracking system based on commercially available off-the-shelf products. In the proposed system, four V2 NoIR Raspberry cameras are connected to various Raspberry Pi boards (Model 3B, 3B+, and 4) as capture stations. In this work, algorithms for clock synchronization, rapid contour extraction, and intrinsic camera calibration are discussed. The size, layout, illumination, and safety of an arena are also addressed, as well as construction issues like non-uniform lighting or noisy reflections. The system's accuracy is sub-centimeter at a frame rate of 100Hz, which is comparable to the performance of the proprietary and commercially available optical tracking systems. These results demonstrate that the proposed solution is feasible and show the correctness of the suggested methodology.
{"title":"Optical tracking system based on COTS components","authors":"Débora N.P. Oliveira, Marcos R. A. Morais, Antonio M. N. Lima","doi":"10.1109/IPAS55744.2022.10053039","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10053039","url":null,"abstract":"This paper deals with the design of an indoor optical tracking system based on commercially available off-the-shelf products. In the proposed system, four V2 NoIR Raspberry cameras are connected to various Raspberry Pi boards (Model 3B, 3B+, and 4) as capture stations. In this work, algorithms for clock synchronization, rapid contour extraction, and intrinsic camera calibration are discussed. The size, layout, illumination, and safety of an arena are also addressed, as well as construction issues like non-uniform lighting or noisy reflections. The system's accuracy is sub-centimeter at a frame rate of 100Hz, which is comparable to the performance of the proprietary and commercially available optical tracking systems. These results demonstrate that the proposed solution is feasible and show the correctness of the suggested methodology.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131355264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10053010
Marinos Louka, Andreas Passos, Antonis Inglezakis, Constantinos Loizou, E. Kaliviotis
Hemostasis is a defence mechanism that prevents blood losses in cases of vessel injuries, and other related disorders. In many cases, patients need to frequently monitor their blood coagulation tendency in order to regulate their medication. In addition, red blood cell aggregation (RBCA) is related to blood inflammation, and it appears elevated in many pathological conditions. Blood coagulation and RBCA can be studied by analysing the dynamic changes of light transmittance though a clotting/aggregating sample, and indeed various works in the literature exploit this approach. In this work, blood coagulation and RBCA are examined by utilising single drops of blood in an inexpensive camera-based microfluidic system, designed for low computational and production cost. Results are compared with a microscopy-camera system, with both setups utilizing the same custom made microchannel. Three image processing algorithms are developed to analyze the averaged light intensity, and the local structural chracteristics of blood, through a binarization and region classification method, using logical operations. The results illustrate the repeatability of the technique and the donor-to-donor variation within the proposed approach. Based on the image processing analysis, the developed coagulation and aggregation indices show great potential of utilisation in an inexpensive and robust point of care device.
{"title":"A microfluidic system, utilising image processing methods, for the detection of blood coagulation and erythrocyte aggregation","authors":"Marinos Louka, Andreas Passos, Antonis Inglezakis, Constantinos Loizou, E. Kaliviotis","doi":"10.1109/IPAS55744.2022.10053010","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10053010","url":null,"abstract":"Hemostasis is a defence mechanism that prevents blood losses in cases of vessel injuries, and other related disorders. In many cases, patients need to frequently monitor their blood coagulation tendency in order to regulate their medication. In addition, red blood cell aggregation (RBCA) is related to blood inflammation, and it appears elevated in many pathological conditions. Blood coagulation and RBCA can be studied by analysing the dynamic changes of light transmittance though a clotting/aggregating sample, and indeed various works in the literature exploit this approach. In this work, blood coagulation and RBCA are examined by utilising single drops of blood in an inexpensive camera-based microfluidic system, designed for low computational and production cost. Results are compared with a microscopy-camera system, with both setups utilizing the same custom made microchannel. Three image processing algorithms are developed to analyze the averaged light intensity, and the local structural chracteristics of blood, through a binarization and region classification method, using logical operations. The results illustrate the repeatability of the technique and the donor-to-donor variation within the proposed approach. Based on the image processing analysis, the developed coagulation and aggregation indices show great potential of utilisation in an inexpensive and robust point of care device.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129837148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052982
E. Vasileva, Z. Ivanovski
This paper presents a convolutional neural network (CNN) architecture for segmenting partially transparent shipping bags in RGB-D images of cluttered scenes containing different packaging items in unstructured configurations. The proposed architecture is optimized for training with a limited number of samples with high variability. The analysis of the results with regard to the input type, network architecture, and lighting conditions, proves that including low-resolution depth information improves the segmentation of objects with similar colors and objects in previously unseen lighting conditions, and the high-resolution color photographs greatly improve the segmentation of details. This motivates the proposed multi-input architecture with early feature fusion in order to fully utilize the benefits of high-resolution photographs and low-resolution depth information. The proposed CNN architecture performs successful segmentation of shipping bags in a cluttered environment among packages and items of different colors and materials with irregular shapes. The CNN provides an improvement in accuracy over well-known semantic segmentation architectures while significantly reducing the required processing time, making it suitable for real-time application.
{"title":"Segmentation of Shipping Bags in RGB-D Images","authors":"E. Vasileva, Z. Ivanovski","doi":"10.1109/IPAS55744.2022.10052982","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052982","url":null,"abstract":"This paper presents a convolutional neural network (CNN) architecture for segmenting partially transparent shipping bags in RGB-D images of cluttered scenes containing different packaging items in unstructured configurations. The proposed architecture is optimized for training with a limited number of samples with high variability. The analysis of the results with regard to the input type, network architecture, and lighting conditions, proves that including low-resolution depth information improves the segmentation of objects with similar colors and objects in previously unseen lighting conditions, and the high-resolution color photographs greatly improve the segmentation of details. This motivates the proposed multi-input architecture with early feature fusion in order to fully utilize the benefits of high-resolution photographs and low-resolution depth information. The proposed CNN architecture performs successful segmentation of shipping bags in a cluttered environment among packages and items of different colors and materials with irregular shapes. The CNN provides an improvement in accuracy over well-known semantic segmentation architectures while significantly reducing the required processing time, making it suitable for real-time application.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"84 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122941461","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10053049
Hassaan A. Qazi
Rendering photorealistic images from the image stylization technique is still considered as a challenging task. In this paper, we compare three recent state-of-the-art approaches. All three algorithms are mainly driven by Convolution Neural Network (CNN) technique. A brief discussion of the selected approaches is followed by some comparisons and results. Both Structural Similarity Index (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS) metrics are used to generate new findings of the methodologies. Finally, subjective analysis is also presented to gauge the efficacy of the algorithms in discussion.
{"title":"A Review of Photorealistic Image Stylization Techniques","authors":"Hassaan A. Qazi","doi":"10.1109/IPAS55744.2022.10053049","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10053049","url":null,"abstract":"Rendering photorealistic images from the image stylization technique is still considered as a challenging task. In this paper, we compare three recent state-of-the-art approaches. All three algorithms are mainly driven by Convolution Neural Network (CNN) technique. A brief discussion of the selected approaches is followed by some comparisons and results. Both Structural Similarity Index (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS) metrics are used to generate new findings of the methodologies. Finally, subjective analysis is also presented to gauge the efficacy of the algorithms in discussion.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121298600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052863
Mouna Zouari Mehdi, A. Benzinou, J. Elleuch, K. Nasreddine, Dhia Ammeri, D. Sellami
Dendritic cells can be seen as a mirror of our immune system. Based on their in virto analysis, biological experts are now able to study the impact of food contaminants on the human immune system. Accordingly, a visual characterization of dendritic cell morphology can provide an indirect estimation of the toxicity. In this paper, we propose an automatic classification of dendritic cells that could serve as a second non-subjective opinion for pathologists. The proposed approach is built on pre-processing steps for segmentation and cell detection in microscopic images. Then, a set of features such as shape descriptors are extracted for cell characterization. At this step, three cell classes are distinctively identified by experts. Nevertheless, a high ambiguity is revealed between cell classes. Possibility theory can offer a realistic framework for making reliable decisions under high ambiguity. It exploits a human natural concept of the implicit use of probability distribution for deciding on the possibility of some assertions in some contexts where a cognitive conflict is observed while interfering existing related postulates, leading to high ambiguity. Based on the consistency concept of Dubois and Prade, a transformation of the probability into a possibility distribution is undertaken. Under possibility paradigm, a further feature selection in the possibility space using the Shapely index. Compared to state-of-the art methods the proposed approach yielded on a real dataset of nearly 630 samples an improvement in terms of the mean precision rate, the Recall rate, and the F1-measure.
{"title":"Human Dendritic Cells Classification based on Possibility Theory","authors":"Mouna Zouari Mehdi, A. Benzinou, J. Elleuch, K. Nasreddine, Dhia Ammeri, D. Sellami","doi":"10.1109/IPAS55744.2022.10052863","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052863","url":null,"abstract":"Dendritic cells can be seen as a mirror of our immune system. Based on their in virto analysis, biological experts are now able to study the impact of food contaminants on the human immune system. Accordingly, a visual characterization of dendritic cell morphology can provide an indirect estimation of the toxicity. In this paper, we propose an automatic classification of dendritic cells that could serve as a second non-subjective opinion for pathologists. The proposed approach is built on pre-processing steps for segmentation and cell detection in microscopic images. Then, a set of features such as shape descriptors are extracted for cell characterization. At this step, three cell classes are distinctively identified by experts. Nevertheless, a high ambiguity is revealed between cell classes. Possibility theory can offer a realistic framework for making reliable decisions under high ambiguity. It exploits a human natural concept of the implicit use of probability distribution for deciding on the possibility of some assertions in some contexts where a cognitive conflict is observed while interfering existing related postulates, leading to high ambiguity. Based on the consistency concept of Dubois and Prade, a transformation of the probability into a possibility distribution is undertaken. Under possibility paradigm, a further feature selection in the possibility space using the Shapely index. Compared to state-of-the art methods the proposed approach yielded on a real dataset of nearly 630 samples an improvement in terms of the mean precision rate, the Recall rate, and the F1-measure.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125860370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10053017
Francis Williams, L. Kuncheva, Juan José Rodríguez Diez, Samuel L. Hennessey
While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both.
{"title":"Combination of Object Tracking and Object Detection for Animal Recognition","authors":"Francis Williams, L. Kuncheva, Juan José Rodríguez Diez, Samuel L. Hennessey","doi":"10.1109/IPAS55744.2022.10053017","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10053017","url":null,"abstract":"While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both.","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122049748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-05DOI: 10.1109/IPAS55744.2022.10052818
Božidar Kelava, M. Vranješ, D. Vranješ, Ž. Lukač
To save transmission, processing and memory resources in Advanced Driver Assistance Systems (ADAS), it is often necessary to reduce the image resolution. Sometimes it is necessary to increase it after the transmission. Both resolution changes involve an image interpolation process. This paper describes implementation for three well-known interpolation methods, nearest neighbour interpolation (NN), bilinear interpolation (BL) and bicubic interpolation (BC), on a real automotive AMV ALPHA platform, using multiple processors on the same System on Chip (SoC). Implementation was done using C programming language and Vision Software Development Kit (VSDK). Specific attention is given to the optimal distribution of tasks to the certain processor. The results have shown that, on the real automotive AMV ALPHA platform, BL interpolation achieves the best trade-off between the quality of interpolated image for the usage in automotive image-processing based algorithms and execution time, especially for the algorithms where the lower frame rate is acceptable (like surround-view, park assist, etc.).
为了节省高级驾驶辅助系统(ADAS)的传输、处理和内存资源，通常需要降低图像分辨率。有时需要在传播后增加它。这两种分辨率变化都涉及图像插值过程。本文介绍了三种著名的插值方法，即最近邻插值(NN)、双线性插值(BL)和双三次插值(BC)，在一个实际的汽车AMV ALPHA平台上，使用同一片上系统(SoC)上的多个处理器实现。使用C语言和visual Software Development Kit (VSDK)实现。特别注意任务的最优分配给特定的处理器。结果表明，在真实的汽车AMV ALPHA平台上，对于基于汽车图像处理的算法，特别是对于可接受较低帧率的算法(如环视、停车辅助等)，BL插值实现了插值图像质量与执行时间之间的最佳权衡。
{"title":"Image Interpolation with Edges Preserving and Implementation on the Real ADAS Platform","authors":"Božidar Kelava, M. Vranješ, D. Vranješ, Ž. Lukač","doi":"10.1109/IPAS55744.2022.10052818","DOIUrl":"https://doi.org/10.1109/IPAS55744.2022.10052818","url":null,"abstract":"To save transmission, processing and memory resources in Advanced Driver Assistance Systems (ADAS), it is often necessary to reduce the image resolution. Sometimes it is necessary to increase it after the transmission. Both resolution changes involve an image interpolation process. This paper describes implementation for three well-known interpolation methods, nearest neighbour interpolation (NN), bilinear interpolation (BL) and bicubic interpolation (BC), on a real automotive AMV ALPHA platform, using multiple processors on the same System on Chip (SoC). Implementation was done using C programming language and Vision Software Development Kit (VSDK). Specific attention is given to the optimal distribution of tasks to the certain processor. The results have shown that, on the real automotive AMV ALPHA platform, BL interpolation achieves the best trade-off between the quality of interpolated image for the usage in automotive image-processing based algorithms and execution time, especially for the algorithms where the lower frame rate is acceptable (like surround-view, park assist, etc.).","PeriodicalId":322228,"journal":{"name":"2022 IEEE 5th International Conference on Image Processing Applications and Systems (IPAS)","volume":"Five 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129741653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：