{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/speedaccuracy-trade-offs-for-modern","title":"Speed/accuracy trade-offs for modern convolutional object detectors","arxiv_id":"1611.10012","date":"2016-11-30","proceeding":"CVPR 2017 7","authors":["Jonathan Huang","Vivek Rathod","Chen Sun","Menglong Zhu","Anoop Korattikara","Alireza Fathi","Ian Fischer","Zbigniew Wojna","Yang song","Sergio Guadarrama","Kevin Murphy"],"abstract":"The goal of this paper is to serve as a guide for selecting a detection\narchitecture that achieves the right speed/memory/accuracy balance for a given\napplication and platform. To this end, we investigate various ways to trade\naccuracy for speed and memory usage in modern convolutional object detection\nsystems. A number of successful systems have been proposed in recent years, but\napples-to-apples comparisons are difficult due to different base feature\nextractors (e.g., VGG, Residual Networks), different default image resolutions,\nas well as different hardware and software platforms. We present a unified\nimplementation of the Faster R-CNN [Ren et al., 2015], R-FCN [Dai et al., 2016]\nand SSD [Liu et al., 2015] systems, which we view as \"meta-architectures\" and\ntrace out the speed/accuracy trade-off curve created by using alternative\nfeature extractors and varying other critical parameters such as image size\nwithin each of these meta-architectures. On one extreme end of this spectrum\nwhere speed and memory are critical, we present a detector that achieves real\ntime speeds and can be deployed on a mobile device. On the opposite end in\nwhich accuracy is critical, we present a detector that achieves\nstate-of-the-art performance measured on the COCO detection task.","url_abs":"http://arxiv.org/abs/1611.10012v3","url_pdf":"http://arxiv.org/pdf/1611.10012v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/IBM/MAX-Object-Detector","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/MobileNetV1_SSD_OpenCV_Caffe","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/MobileNet_SSD_OpenCV_TensorFlow","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/TensorFlow_Lite_RPi_32-bits","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/TensorFlow_Lite_RPi_64-bits","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/TensorFlow_Lite_SSD_Jetson-Nano","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/TensorFlow_Lite_SSD_RPi_32-bits","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/Qengineering/TensorFlow_Lite_SSD_RPi_64-bits","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/jiportilla/edge-max-object-detector","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/rajatashhpa/1","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/sdkchris/Exercise-Sheet-1","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/tensorflow/models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/tensorflow/models/tree/master/research/object_detection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"speedaccuracy-trade-offs-for-modern","repo_url":"https://github.com/yuhsijen/Object-Detector","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"faster-r-cnn","method_name":"Faster R-CNN"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"non-maximum-suppression","method_name":"Non Maximum Suppression"},{"method_slug":"position-sensitive-roi-pooling","method_name":"Position-Sensitive RoI Pooling"},{"method_slug":"r-fcn","method_name":"R-FCN"},{"method_slug":"rpn","method_name":"RPN"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"roipool","method_name":"RoIPool"},{"method_slug":"speed","method_name":"SPEED"},{"method_slug":"ssd","method_name":"SSD"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1611.10012","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}