{"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/attention-based-context-aggregation-network","title":"Attention-based Context Aggregation Network for Monocular Depth Estimation","arxiv_id":"1901.10137","date":"2019-01-29","proceeding":null,"authors":["Yuru Chen","Haitao Zhao","Zhengwei Hu"],"abstract":"Depth estimation is a traditional computer vision task, which plays a crucial\nrole in understanding 3D scene geometry. Recently,\ndeep-convolutional-neural-networks based methods have achieved promising\nresults in the monocular depth estimation field. Specifically, the framework\nthat combines the multi-scale features extracted by the dilated convolution\nbased block (atrous spatial pyramid pooling, ASPP) has gained the significant\nimprovement in the dense labeling task. However, the discretized and predefined\ndilation rates cannot capture the continuous context information that differs\nin diverse scenes and easily introduce the grid artifacts in depth estimation.\nIn this paper, we propose an attention-based context aggregation network (ACAN)\nto tackle these difficulties. Based on the self-attention model, ACAN\nadaptively learns the task-specific similarities between pixels to model the\ncontext information. First, we recast the monocular depth estimation as a dense\nlabeling multi-class classification problem. Then we propose a soft ordinal\ninference to transform the predicted probabilities to continuous depth values,\nwhich can reduce the discretization error (about 1% decrease in RMSE). Second,\nthe proposed ACAN aggregates both the image-level and pixel-level context\ninformation for depth estimation, where the former expresses the statistical\ncharacteristic of the whole image and the latter extracts the long-range\nspatial dependencies for each pixel. Third, for further reducing the\ninconsistency between the RGB image and depth map, we construct an attention\nloss to minimize their information entropy. We evaluate on public monocular\ndepth-estimation benchmark datasets (including NYU Depth V2, KITTI). The\nexperiments demonstrate the superiority of our proposed ACAN and achieve the\ncompetitive results with the state of the arts.","url_abs":"http://arxiv.org/abs/1901.10137v1","url_pdf":"http://arxiv.org/pdf/1901.10137v1.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":"attention-based-context-aggregation-network","repo_url":"https://github.com/miraiaroha/ACAN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"depth-estimation","task_name":"Depth Estimation"},{"task_slug":"monocular-depth-estimation","task_name":"Monocular Depth Estimation"},{"task_slug":"multi-class-classification","task_name":"Multi-class Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/monocular-depth-estimation-on-nyu-depth-v2","task":"Monocular Depth Estimation","dataset":"NYU-Depth V2","model":"ACAN","rank_in_archive_order":64,"of":85,"metrics":{"RMSE":"0.496"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.10137","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1901.10137"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/miraiaroha/ACAN","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":12},"by_repo_kind":{"official":{"samples":12,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"f1f5ec5fdff029e7","entry":"NormalDist","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/losses.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/losses.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f1f5ec5fdff029e7"}},{"code_sha256_prefix":"c80c792bc496ed80","entry":"continuous2discrete","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/model.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c80c792bc496ed80"}},{"code_sha256_prefix":"6c3dd8fb65a40c15","entry":"create_datasets","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/dataloaders/get_datasets.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/dataloaders/get_datasets.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6c3dd8fb65a40c15"}},{"code_sha256_prefix":"1b4a913088f7e432","entry":"create_lossfunc","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/get_lossfunc.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/get_lossfunc.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1b4a913088f7e432"}},{"code_sha256_prefix":"7f991c24c3f646bc","entry":"create_network","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/get_network.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/get_network.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7f991c24c3f646bc"}},{"code_sha256_prefix":"d07cbe27712d5d01","entry":"create_optimizer","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/creator/optimizer.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/creator/optimizer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d07cbe27712d5d01"}},{"code_sha256_prefix":"419d8a3c9480d263","entry":"create_params","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/creator/params.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/creator/params.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"419d8a3c9480d263"}},{"code_sha256_prefix":"c60cc48bb993e02b","entry":"create_scheduler","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/creator/scheduler.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/creator/scheduler.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c60cc48bb993e02b"}},{"code_sha256_prefix":"808145c33a4f073c","entry":"discrete2continuous","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/model.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"808145c33a4f073c"}},{"code_sha256_prefix":"8049f1ba083fcffa","entry":"get_params","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/creator/params.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/creator/params.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8049f1ba083fcffa"}},{"code_sha256_prefix":"b66f1ba79bd18978","entry":"img_loader","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/dataloaders/dataloader.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/dataloaders/dataloader.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b66f1ba79bd18978"}},{"code_sha256_prefix":"07440157233b90f7","entry":"make_decoder","repo":"miraiaroha/ACAN","repo_kind":"official","path":"code/models/model.py","file_url":"https://github.com/miraiaroha/ACAN/blob/HEAD/code/models/model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"07440157233b90f7"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}