Research map / Detection and segmentation / CNN features
Object Detection from Scratch with Deep Supervision
Zhiqiang Shen, Zhuang Liu, Jianguo Li, Yu–Gang Jiang, Yurong Chen, Xiangyang Xue
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add stem block (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): 74.5 -> 77.3 | Improved |
| add compression factor theta=1 in transition layers (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): +2.9% mAP | Improved |
| add wider bottleneck layers (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): +4.1% mAP | Improved |
| add more channels in first conv layer (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): +1.1% mAP | Improved |
| increase growth rate from 16 to 48 (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): +4.8% mAP | Improved |
| add transition w/o pooling layer (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): +1.7% | Improved |
| add dense prediction structure (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): 77.3 -> 77.7 | Improved |
| add deep-scale supervision (DSS) module (ablation) | DSOD300 · VOC 2007 test (07+12, from scratch) | mAP (%): 77.7 -> 79.1 | Improved |
| add DSS module and dense prediction to SSD (ablation) | SSD from scratch (VGGNet) · VOC 2007 test (07+12, from scratch) | mAP (%): 70.4 -> 77.4 | Improved |
| add batch norm to DSS SSD (ablation) | SSD+DP+DSS · VOC 2007 test (07+12, from scratch) | mAP (%): 74.2 -> 77.4 | Improved |
| use deep supervision via dense blocks instead of plain VGG, from scratch | SSD300S · VOC 2007 test | mAP (%): 69.6 -> 77.7 | Improved |
| add BN to two-stage detector trained from scratch (ablation) | Faster R-CNN VGG16 RoI Pool · VOC 2007 test | mAP (%): 6.4 -> 37.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- R-FCN — R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016) · better on 1 of 2 shared results
- Faster RCNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DSOD: Learning Deeply Supervised Object Detectors From Scratch (ICCV 2017 · similarity 0.98)
- ScratchDet: Training Single-Shot Object Detectors From Scratch (CVPR 2019 · similarity 0.91)
- Deep Learning for Generic Object Detection: A Survey (IJCV 2019 · similarity 0.91)
- Salient Object Detection in the Deep Learning Era: An In-Depth Survey (TPAMI 2021 · similarity 0.90)
- Learning Rich Features at High-Speed for Single-Shot Object Detection (ICCV 2019 · similarity 0.89)
- LEOD: Label-Efficient Object Detection for Event Cameras (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
