Dynamic multi-scale loss optimization for object detection
نویسندگان
چکیده
With the continuous improvement of deep object detectors via advanced model architectures, imbalance problems in training process have received more attention. It is a common paradigm detection frameworks to perform multi-scale detection. However, each scale treated equally during training. In this paper, we carefully study objective detector We argue that loss level neither important nor independent. Different from existing solutions setting multi-task weights, dynamically optimize weight process. Specifically, propose an Adaptive Variance Weighting (AVW) balance according statistical variance. Then develop novel Reinforcement Learning Optimization (RLO) decide weighting scheme probabilistically makes better utilization without extra computational complexity and learnable parameters for backpropagation. Without bells whistles, proposed method improves ATSS by 0.9 AP on MS COCO benchmark. And it achieves 82.1 mAP Pascal VOC 2007 test set, which outperforms other reinforcement-learning-based methods.
منابع مشابه
Multi-scale Location-aware Kernel Representation for Object Detection
Although Faster R-CNN and its variants have shown promising performance in object detection, they only exploit simple first-order representation of object proposals for final classification and regression. Recent classification methods demonstrate that the integration of highorder statistics into deep convolutional neural networks can achieve impressive improvement, but their goal is to model w...
متن کاملMSDNN: Multi-Scale Deep Neural Network for Salient Object Detection
Salient object detection is a fundamental problem and has been received a great deal of attentions in computer vision. Recently deep learning model became a powerful tool for image feature extraction. In this paper, we propose a multi-scale deep neural network (MSDNN) for salient object detection. The proposed model first extracts global high-level features and context information over the whol...
متن کاملMulti-scale volumes for deep object detection and localization
This study aims to analyze the benefits of improved multiscale reasoning for object detection and localization with deep convolutional neural networks. To that end, an efficient and general object detection framework which operates on scale volumes of a deep feature pyramid is proposed. In contrast to the proposed approach, most current state-of-the-art object detectors operate on a single-scal...
متن کاملMulti-scale Cortical Keypoint Representation for Attention and Object Detection
Keypoints (junctions) provide important information for focus-of-attention (FoA) and object categorization/recognition. In this paper we analyze the multi-scale keypoint representation, obtained by applying a linear and quasi-continuous scaling to an optimized model of cortical end-stopped cells, in order to study its importance and possibilities for developing a visual, cortical architecture. ...
متن کاملMulti-Objective Neural Network Optimization for Visual Object Detection
In real-time computer vision, there is a need for classifiers that detect patterns fast and reliably. We apply multi-objective optimization (MOO) to the design of feed-forward neural networks for real-world object recognition tasks, where computational complexity and accuracy define partially conflicting objectives. Evolutionary structure optimization and pruning are compared for the adaptation...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Multimedia Tools and Applications
سال: 2022
ISSN: ['1380-7501', '1573-7721']
DOI: https://doi.org/10.1007/s11042-022-13164-9