{"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/an-analysis-of-scale-invariance-in-object-1","title":"An Analysis of Scale Invariance in Object Detection - SNIP","arxiv_id":"1711.08189","date":"2017-11-22","proceeding":null,"authors":["Bharat Singh","Larry S. Davis"],"abstract":"An analysis of different techniques for recognizing and detecting objects\nunder extreme scale variation is presented. Scale specific and scale invariant\ndesign of detectors are compared by training them with different configurations\nof input data. By evaluating the performance of different network architectures\nfor classifying small objects on ImageNet, we show that CNNs are not robust to\nchanges in scale. Based on this analysis, we propose to train and test\ndetectors on the same scales of an image-pyramid. Since small and large objects\nare difficult to recognize at smaller and larger scales respectively, we\npresent a novel training scheme called Scale Normalization for Image Pyramids\n(SNIP) which selectively back-propagates the gradients of object instances of\ndifferent sizes as a function of the image scale. On the COCO dataset, our\nsingle model performance is 45.7% and an ensemble of 3 networks obtains an mAP\nof 48.3%. We use off-the-shelf ImageNet-1000 pre-trained models and only train\nwith bounding box supervision. Our submission won the Best Student Entry in the\nCOCO 2017 challenge. Code will be made available at\n\\url{http://bit.ly/2yXVg4c}.","url_abs":"http://arxiv.org/abs/1711.08189v2","url_pdf":"http://arxiv.org/pdf/1711.08189v2.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":[],"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":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dpn","method_name":"DPN"},{"method_slug":"dpn-block","method_name":"DPN Block"},{"method_slug":"deformable-convolution","method_name":"Deformable Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"grouped-convolution","method_name":"Grouped Convolution"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"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":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"snip","method_name":"SNIP"},{"method_slug":"soft-nms","method_name":"Soft-NMS"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[{"slug":"snip","name":"SNIP","full_name":"SNIP"}],"results":[{"leaderboard":"/sota/object-detection-on-coco","task":"Object Detection","dataset":"COCO test-dev","model":"D-RFCN + SNIP (DPN-98 with flip, multi-scale)","rank_in_archive_order":137,"of":225,"metrics":{"AP50":"67.3","AP75":"51.1","APL":"57.1","APM":"48.8","APS":"29.3","box mAP":"45.7"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-coco","task":"Object Detection","dataset":"COCO test-dev","model":"D-RFCN + SNIP (ResNet-101, multi-scale)","rank_in_archive_order":159,"of":225,"metrics":{"AP50":"65.5","AP75":"48.4","APL":"54.9","APM":"46.5","APS":"27.2","box mAP":"43.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.08189","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}