Pub Date : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293926
H. Verma, Siddharth Lotia, Ashutosh Kumar Singh
Various recent advancements in deep learning models have greatly boosted the performance of semantic pattern recognition using images. Various state estimation of an individual like emotional state and other certain character features or traits can be estimated from the facial images. With this motivation, in this work we are attempting to infer criminal tendency or (crime prediction/detection) from facial images by using the learning capabilities of various deep learning architectures. More precisely two type of deep learning models we have used in this study: standard convolutional neural network(CNN) architecture and pre-trained CNN architectures, namely VGG-16, VGG-19, and Incep-tionV3. We have done a performance comparative analysis among these models for efficiently capturing criminal traits from a human face. The efficacy of the above deep learning models was evaluated on a public database, National Institute of Standards and Technology (NIST). To avoid any discrepancies, we have only used male images in this work. It was found that VGG CNN models are best performing models, especially in a limited data scenario producing the classification accuracy of 99.5% in identifying criminal faces.
{"title":"Convolutional Neural Network Based Criminal Detection","authors":"H. Verma, Siddharth Lotia, Ashutosh Kumar Singh","doi":"10.1109/TENCON50793.2020.9293926","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293926","url":null,"abstract":"Various recent advancements in deep learning models have greatly boosted the performance of semantic pattern recognition using images. Various state estimation of an individual like emotional state and other certain character features or traits can be estimated from the facial images. With this motivation, in this work we are attempting to infer criminal tendency or (crime prediction/detection) from facial images by using the learning capabilities of various deep learning architectures. More precisely two type of deep learning models we have used in this study: standard convolutional neural network(CNN) architecture and pre-trained CNN architectures, namely VGG-16, VGG-19, and Incep-tionV3. We have done a performance comparative analysis among these models for efficiently capturing criminal traits from a human face. The efficacy of the above deep learning models was evaluated on a public database, National Institute of Standards and Technology (NIST). To avoid any discrepancies, we have only used male images in this work. It was found that VGG CNN models are best performing models, especially in a limited data scenario producing the classification accuracy of 99.5% in identifying criminal faces.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129164300","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293786
Devi T. Avalokita, Tessya Rismonita, A. Handayani, A. W. Setiawan
Gestational age (GA) monitoring from fetal ultrasound imaging is one method to observe pre-birth risk factors and to prepare early treatment for neonatal problems. There are several parameters in an ultrasound image that can be used to estimate GA, one of which is the fetal head circumference (HC). However, fetal HC measurement is prone to error since it relies on manual annotation by sonographer or obstetrician. This research aims to design an algorithm to automatically calculate the fetal HC based on optimized ellipse fitting on a localized region of interest (RoI) previously defined as fetal head candidate area. Our optimization method consists of pre-processing steps to exclude noise within the RoI and to select the optimum representation of fetal head pixels to be processed by the ellipse fitting algorithm. We managed to perform ellipse fitting on 699 and 141 ultrasound images representing respectively the second and third trimester pregnancies; with the average dice similarity coefficient (DSC) of 95.27%±6.25%, hausdorff distance (HD) of 3.51 mm±5.54 mm, a difference in fetal HC (DF) of -3.42 mm±13.66 mm, and an absolute difference in fetal HC (ADF) of 6.53 mm±12.5 mm. The results demonstrated that the presented method performed comparably to other systems published in the literature. Moreover, our results represent an evaluation of a significantly larger number of data compared to most of the previous works.
{"title":"Automatic Fetal Head Circumference Measurement in 2D Ultrasound Images Based On Optimized Fast Ellipse Fitting","authors":"Devi T. Avalokita, Tessya Rismonita, A. Handayani, A. W. Setiawan","doi":"10.1109/TENCON50793.2020.9293786","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293786","url":null,"abstract":"Gestational age (GA) monitoring from fetal ultrasound imaging is one method to observe pre-birth risk factors and to prepare early treatment for neonatal problems. There are several parameters in an ultrasound image that can be used to estimate GA, one of which is the fetal head circumference (HC). However, fetal HC measurement is prone to error since it relies on manual annotation by sonographer or obstetrician. This research aims to design an algorithm to automatically calculate the fetal HC based on optimized ellipse fitting on a localized region of interest (RoI) previously defined as fetal head candidate area. Our optimization method consists of pre-processing steps to exclude noise within the RoI and to select the optimum representation of fetal head pixels to be processed by the ellipse fitting algorithm. We managed to perform ellipse fitting on 699 and 141 ultrasound images representing respectively the second and third trimester pregnancies; with the average dice similarity coefficient (DSC) of 95.27%±6.25%, hausdorff distance (HD) of 3.51 mm±5.54 mm, a difference in fetal HC (DF) of -3.42 mm±13.66 mm, and an absolute difference in fetal HC (ADF) of 6.53 mm±12.5 mm. The results demonstrated that the presented method performed comparably to other systems published in the literature. Moreover, our results represent an evaluation of a significantly larger number of data compared to most of the previous works.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123569313","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293869
Desi Budiastuti, Ardine Khairunisa Ilyas, E. Tjipto Rahardjo
The antenna proposed in this study is a wearable microstrip patch antenna that utilizes jeans (permittivity: 1.77) as its substrate for GPS application. Tests shows that the antenna has frequency range of 1.57 – 1.61 GHz. Resonant frequency of the antenna is 1.595 GHz, with return loss value of -14.18 dB. The antenna achieved its desired specification with truncated edge, quarter wave transformator, and slot utilization. The antenna is safe to be used on thigh, chest, and arm as simulation shows that SAR value of the antenna is under the maximum standard allowed. However, when the antenna is moved further away from the phantom, the axial ratio value decreases and goes > 3 dB when antenna is placed over the distance recommendation.
{"title":"Design and Assembly of Textile Microstrip Antenna for Global Positioning System Application","authors":"Desi Budiastuti, Ardine Khairunisa Ilyas, E. Tjipto Rahardjo","doi":"10.1109/TENCON50793.2020.9293869","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293869","url":null,"abstract":"The antenna proposed in this study is a wearable microstrip patch antenna that utilizes jeans (permittivity: 1.77) as its substrate for GPS application. Tests shows that the antenna has frequency range of 1.57 – 1.61 GHz. Resonant frequency of the antenna is 1.595 GHz, with return loss value of -14.18 dB. The antenna achieved its desired specification with truncated edge, quarter wave transformator, and slot utilization. The antenna is safe to be used on thigh, chest, and arm as simulation shows that SAR value of the antenna is under the maximum standard allowed. However, when the antenna is moved further away from the phantom, the axial ratio value decreases and goes > 3 dB when antenna is placed over the distance recommendation.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"61 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120926308","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293918
Arjun Suresh, B. N. Reddy, C. Madhavi
The proliferation of IoT devices in recent years has resulted in an exponential increase in data being transmitted over the internet. The traffic is slated for further increase in the coming years and will result in excessive network congestion and high latency. To alleviate this problem, an alternate approach needs to be considered. A prominent option would be to move the computing domain to the edge device. This option is constrained due to reduced computing, storage and power available on the edge. A novel approach combining both software and hardware solutions is required to perform analytics at the edge. This paper proposes an architecture for analysing data on the edge, combining hardware and software solutions. The proposed methodology explores machine learning algorithms for edge computing combined with the use of hardware accelerators to achieve truly intelligent edge devices. A qualitative and quantitative comparison of performance of various algorithms on CPU, GPU, FPGA platforms is carried out. A machine learning model for predicting Remaining Useful Life (RUL) for a multivariate time series dataset is developed and its deployment on the edge is discussed. The results of the experiments carried out are promising and hold potential for further research.
{"title":"Hardware Accelerators for Edge Enabled Machine Learning","authors":"Arjun Suresh, B. N. Reddy, C. Madhavi","doi":"10.1109/TENCON50793.2020.9293918","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293918","url":null,"abstract":"The proliferation of IoT devices in recent years has resulted in an exponential increase in data being transmitted over the internet. The traffic is slated for further increase in the coming years and will result in excessive network congestion and high latency. To alleviate this problem, an alternate approach needs to be considered. A prominent option would be to move the computing domain to the edge device. This option is constrained due to reduced computing, storage and power available on the edge. A novel approach combining both software and hardware solutions is required to perform analytics at the edge. This paper proposes an architecture for analysing data on the edge, combining hardware and software solutions. The proposed methodology explores machine learning algorithms for edge computing combined with the use of hardware accelerators to achieve truly intelligent edge devices. A qualitative and quantitative comparison of performance of various algorithms on CPU, GPU, FPGA platforms is carried out. A machine learning model for predicting Remaining Useful Life (RUL) for a multivariate time series dataset is developed and its deployment on the edge is discussed. The results of the experiments carried out are promising and hold potential for further research.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116264818","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293708
Sango Nagamoto, M. Omiya
The commercial service of fifth generation mobile communication system started in the spring of 2020 in Japan. This paper discusses a highly precise prediction of 28 GHz millimeter wave indoor propagation characteristics in an office environment by using a large-scale electro-magnetic field simulation based on the finite difference time domain technique. The computer simulations are carried out using the high-performance computer system operated in Information Initiative Center, Hokkaido University. They give us a detail of electromagnetic field distributions in an FDTD problem space including targets at once although they require a lot of computer resources and a long running time in general. The paper compares calculated path loss model parameters such as path loss exponents, shadow factors and cross-polarization discriminations in the LOS environment with the measured ones demonstrated by the other research groups to confirm the effectiveness of numerical results and the accurate prediction of path loss model parameters.
{"title":"Highly Precise Prediction of 28 GHz Indoor Radio Wave Propagation Characteristics in an Office Environment for Design of 5G Wireless Networks","authors":"Sango Nagamoto, M. Omiya","doi":"10.1109/TENCON50793.2020.9293708","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293708","url":null,"abstract":"The commercial service of fifth generation mobile communication system started in the spring of 2020 in Japan. This paper discusses a highly precise prediction of 28 GHz millimeter wave indoor propagation characteristics in an office environment by using a large-scale electro-magnetic field simulation based on the finite difference time domain technique. The computer simulations are carried out using the high-performance computer system operated in Information Initiative Center, Hokkaido University. They give us a detail of electromagnetic field distributions in an FDTD problem space including targets at once although they require a lot of computer resources and a long running time in general. The paper compares calculated path loss model parameters such as path loss exponents, shadow factors and cross-polarization discriminations in the LOS environment with the measured ones demonstrated by the other research groups to confirm the effectiveness of numerical results and the accurate prediction of path loss model parameters.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127770372","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293695
M. Kaneda, Kazumasa Ariyoshi, S. Matsumoto
This paper investigate instability in device characteristics related to the hot carrier effect, Negative Bias Temperature Instability (NBTI) and Positive Bias Temperature (PBTI) under DC stress for n- and p-channel thin-film Silicon on Insulator (SOI) power MOSFET at high temperature. The threshold voltage shift increases as the temperature rises due to PBTI for n-MOSFET and NBTI for p-MOSFET. Drain Avalanche Hot Carrier (DAHC) occurs when the gate stress voltage is near the threshold voltage and Channel Hot Carrier (CHC) occurs when the gate voltage is high. The threshold voltage shift and the degradation rate of on-resistance of the n-MOSFET is larger than that of the p-MOSFET due to the difference in the impact ionization coefficient between electrons and holes.
{"title":"Comparisons of instability in device characteristics at high temperature for thin-film SOI power n- and p- channel MOSFETs","authors":"M. Kaneda, Kazumasa Ariyoshi, S. Matsumoto","doi":"10.1109/TENCON50793.2020.9293695","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293695","url":null,"abstract":"This paper investigate instability in device characteristics related to the hot carrier effect, Negative Bias Temperature Instability (NBTI) and Positive Bias Temperature (PBTI) under DC stress for n- and p-channel thin-film Silicon on Insulator (SOI) power MOSFET at high temperature. The threshold voltage shift increases as the temperature rises due to PBTI for n-MOSFET and NBTI for p-MOSFET. Drain Avalanche Hot Carrier (DAHC) occurs when the gate stress voltage is near the threshold voltage and Channel Hot Carrier (CHC) occurs when the gate voltage is high. The threshold voltage shift and the degradation rate of on-resistance of the n-MOSFET is larger than that of the p-MOSFET due to the difference in the impact ionization coefficient between electrons and holes.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125791621","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293935
A. Srisuphab, N. Kaakkurivaara, P. Silapachote, Kitipong Tangkit, Ponthep Meunpong, T. Sunetnanta
Protecting and increasing worldwide green space have been an international effort. Individuals and organizations are encouraged to plant urban trees and to get involved in many reforestation and restoration projects. Offsetting these much needed plans to save the forests is illegal logging. Trees that have grown for many years, some are protected resources inside restricted areas, are felled and the wood is smuggled. Watching for these illegal activities is very difficult and also very dangerous. It is quite impossible for rangers to patrol every entry and exit point of forests that cover thousands of squared kilometers. Applying Internet of Things technology to ecological forestry, we are proposing integrating sound acquisition networks and acoustic signal analyzers to enhance the robustness of an already successful camera-based surveillance solution that is also equipped with a global positioning system tracker. Our listener devices record sounds of the forest and periodically send it to a cloud storage over cellular networks. The device is affordable, the system is small and portable, and the network is flexibly extensible. From the data, acoustic features are extracted and visualized. The Mel-frequency cepstral coefficients of the signals have exhibited promising distinctiveness for detection of illegal chainsaw activities in the wild.
{"title":"Illegal Logging Listeners Using IoT Networks","authors":"A. Srisuphab, N. Kaakkurivaara, P. Silapachote, Kitipong Tangkit, Ponthep Meunpong, T. Sunetnanta","doi":"10.1109/TENCON50793.2020.9293935","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293935","url":null,"abstract":"Protecting and increasing worldwide green space have been an international effort. Individuals and organizations are encouraged to plant urban trees and to get involved in many reforestation and restoration projects. Offsetting these much needed plans to save the forests is illegal logging. Trees that have grown for many years, some are protected resources inside restricted areas, are felled and the wood is smuggled. Watching for these illegal activities is very difficult and also very dangerous. It is quite impossible for rangers to patrol every entry and exit point of forests that cover thousands of squared kilometers. Applying Internet of Things technology to ecological forestry, we are proposing integrating sound acquisition networks and acoustic signal analyzers to enhance the robustness of an already successful camera-based surveillance solution that is also equipped with a global positioning system tracker. Our listener devices record sounds of the forest and periodically send it to a cloud storage over cellular networks. The device is affordable, the system is small and portable, and the network is flexibly extensible. From the data, acoustic features are extracted and visualized. The Mel-frequency cepstral coefficients of the signals have exhibited promising distinctiveness for detection of illegal chainsaw activities in the wild.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132475178","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293867
Chao Liu, Gen Li, Yu Huang, Xiaolong Zhang, Yuanlong Xie, Jie Meng, Liquan Jiang
global localization is essential for pose initialization and pose recovery. However, for the lack of prior information, global localization is always unreliable and time consuming, especially in featureless and dynamic industry environment. To alleviate the negative influence of such environment, this paper uses reflector as landmarks. Then, several maps including labeled occupancy grid map and multi-resolution likelihood field are proposed to model the positions of landmarks as well as ordinary obstacles. Furthermore, a branch and bound method is employed to achieve fast global search based on those proposed maps. Through experiments in a real industry application, the reliability and efficiency of our proposed global localization method is verified.
{"title":"Fast and Reliable Global Localization Using Reflector Landmarks","authors":"Chao Liu, Gen Li, Yu Huang, Xiaolong Zhang, Yuanlong Xie, Jie Meng, Liquan Jiang","doi":"10.1109/TENCON50793.2020.9293867","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293867","url":null,"abstract":"global localization is essential for pose initialization and pose recovery. However, for the lack of prior information, global localization is always unreliable and time consuming, especially in featureless and dynamic industry environment. To alleviate the negative influence of such environment, this paper uses reflector as landmarks. Then, several maps including labeled occupancy grid map and multi-resolution likelihood field are proposed to model the positions of landmarks as well as ordinary obstacles. Furthermore, a branch and bound method is employed to achieve fast global search based on those proposed maps. Through experiments in a real industry application, the reliability and efficiency of our proposed global localization method is verified.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130448736","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293732
R. Banaeeyan, M. F. A. Fauzi, Wei Chen, Debbie Knight, H. Hampel, W. Frankel, M. Gürcan
Tumor buddings (TB), a special formation of cancerous cells that bud from the tumor front, are fast becoming the key indicator in modern clinical applications where they play a significant role in prognostic and evaluation of colorectal cancers in histopathological images. Recently, computational methods have been rapidly evolving in the domain of digital pathology, yet the literature lacks computerized approaches to automate the localization and segmentation of TBs in hematoxylin and eosin (H&E)-stained images. This research addresses this very challenging task of tumor budding detection in H&E images by presenting different deep learning architectures designed for semantic segmentation. The proposed design for a new Convolutional Neural Network (CNN) incorporates convolution filters with different factors of dilations. Multiple experiments based on a newly constructed colorectal cancer histopathological image collection provided promising performances. The best average intersection over union (IOU) for TB of 0.11, IOU for non-TB of 0.86, mean IOU of 0.49 and weighted IOU of 0.83 were observed.
{"title":"Tumor Budding Detection in H&E-Stained Images Using Deep Semantic Learning","authors":"R. Banaeeyan, M. F. A. Fauzi, Wei Chen, Debbie Knight, H. Hampel, W. Frankel, M. Gürcan","doi":"10.1109/TENCON50793.2020.9293732","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293732","url":null,"abstract":"Tumor buddings (TB), a special formation of cancerous cells that bud from the tumor front, are fast becoming the key indicator in modern clinical applications where they play a significant role in prognostic and evaluation of colorectal cancers in histopathological images. Recently, computational methods have been rapidly evolving in the domain of digital pathology, yet the literature lacks computerized approaches to automate the localization and segmentation of TBs in hematoxylin and eosin (H&E)-stained images. This research addresses this very challenging task of tumor budding detection in H&E images by presenting different deep learning architectures designed for semantic segmentation. The proposed design for a new Convolutional Neural Network (CNN) incorporates convolution filters with different factors of dilations. Multiple experiments based on a newly constructed colorectal cancer histopathological image collection provided promising performances. The best average intersection over union (IOU) for TB of 0.11, IOU for non-TB of 0.86, mean IOU of 0.49 and weighted IOU of 0.83 were observed.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134128085","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 : 2020-11-16DOI: 10.1109/TENCON50793.2020.9293825
S. Yadav, K. Sarawadekar
The guided image filter (GIF) technique is used for haze removal. It reduces the gradient reversal artifact as well as preserves the edge information precisely in smooth (flat) region only. However, it fails to avoid halo artifact and edge-smoothing effect in sharp regions. So, to mitigate this problem, we propose an adaptive haze removal algorithm using a steering kernel-based guided image filter (SKGIF). Steering kernel determines the edge-direction in guidance image more adequately. The edge-direction is an essential feature of guidance image, and it helps to determine more edge-preserving information in flat as well as sharp regions. Experimental outcomes on different hazy images prove the effectiveness of the proposed algorithm.
{"title":"Steering Kernel-Based Guided Image Filter for Single Image Dehazing","authors":"S. Yadav, K. Sarawadekar","doi":"10.1109/TENCON50793.2020.9293825","DOIUrl":"https://doi.org/10.1109/TENCON50793.2020.9293825","url":null,"abstract":"The guided image filter (GIF) technique is used for haze removal. It reduces the gradient reversal artifact as well as preserves the edge information precisely in smooth (flat) region only. However, it fails to avoid halo artifact and edge-smoothing effect in sharp regions. So, to mitigate this problem, we propose an adaptive haze removal algorithm using a steering kernel-based guided image filter (SKGIF). Steering kernel determines the edge-direction in guidance image more adequately. The edge-direction is an essential feature of guidance image, and it helps to determine more edge-preserving information in flat as well as sharp regions. Experimental outcomes on different hazy images prove the effectiveness of the proposed algorithm.","PeriodicalId":283131,"journal":{"name":"2020 IEEE REGION 10 CONFERENCE (TENCON)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133869012","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：