Projects per year
Abstract
Major advancements have been made in the field of object
detection and segmentation recently. However, when it comes to rare
categories, the state-of-the-art methods fail to detect them, resulting
in a significant performance gap between rare and frequent categories.
In this paper, we identify that Sigmoid or Softmax functions used in
deep detectors are a major reason for low performance and are suboptimal
for long-tailed detection and segmentation. To address this, we
develop a Gumbel Optimized Loss (GOL), for long-tailed detection and
segmentation. It aligns with the Gumbel distribution of rare classes in
imbalanced datasets, considering the fact that most classes in long-tailed
detection have low expected probability. The proposed GOL significantly
outperforms the best state-of-the-art method by 1.1% on AP, and boosts
the overall segmentation by 9.0% and detection by 8.0%, particularly
improving detection of rare classes by 20.3%, compared to Mask-RCNN,
on LVIS dataset. Code available at: https://github.com/kostas1515/
GOL.
detection and segmentation recently. However, when it comes to rare
categories, the state-of-the-art methods fail to detect them, resulting
in a significant performance gap between rare and frequent categories.
In this paper, we identify that Sigmoid or Softmax functions used in
deep detectors are a major reason for low performance and are suboptimal
for long-tailed detection and segmentation. To address this, we
develop a Gumbel Optimized Loss (GOL), for long-tailed detection and
segmentation. It aligns with the Gumbel distribution of rare classes in
imbalanced datasets, considering the fact that most classes in long-tailed
detection have low expected probability. The proposed GOL significantly
outperforms the best state-of-the-art method by 1.1% on AP, and boosts
the overall segmentation by 9.0% and detection by 8.0%, particularly
improving detection of rare classes by 20.3%, compared to Mask-RCNN,
on LVIS dataset. Code available at: https://github.com/kostas1515/
GOL.
| Original language | English |
|---|---|
| Title of host publication | European Conference on Computer Vision (ECCV) 2022 |
| Publication status | Accepted/In press - 2022 |
Fingerprint
Dive into the research topics of 'Long-tailed Instance Segmentation using Gumbel Optimized Loss'. Together they form a unique fingerprint.Projects
- 1 Finished
-
ViTac: Visual-Tactile Synergy for Handling Flexible Materials
Luo, S. (Primary Investigator)
EPSRC Engineering and Physical Sciences Research Council
17/12/2021 → 11/10/2024
Project: Research