Visual Detector Compression via Location-Aware Discriminant Analysis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lan, Qizhen, Choi, Jung Im, Tian, Qing
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916961057767424
author Lan, Qizhen
Choi, Jung Im
Tian, Qing
author_facet Lan, Qizhen
Choi, Jung Im
Tian, Qing
contents Deep neural networks are powerful, yet their high complexity greatly limits their potential to be deployed on billions of resource-constrained edge devices. Pruning is a crucial network compression technique, yet most existing methods focus on classification models, with limited attention to detection. Even among those addressing detection, there is a lack of utilization of essential localization information. Also, many pruning methods passively rely on pre-trained models, in which useful and useless components are intertwined, making it difficult to remove the latter without harming the former at the neuron/filter level. To address the above issues, in this paper, we propose a proactive detection-discriminants-based network compression approach for deep visual detectors, which alternates between two steps: (1) maximizing and compressing detection-related discriminants and aligning them with a subset of neurons/filters immediately before the detection head, and (2) tracing the detection-related discriminating power across the layers and discarding features of lower importance. Object location information is exploited in both steps. Extensive experiments, employing four advanced detection models and four state-of-the-art competing methods on the KITTI and COCO datasets, highlight the superiority of our approach. Remarkably, our compressed models can even beat the original base models with a substantial reduction in complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2509_17968
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Visual Detector Compression via Location-Aware Discriminant Analysis
Lan, Qizhen
Choi, Jung Im
Tian, Qing
Computer Vision and Pattern Recognition
Deep neural networks are powerful, yet their high complexity greatly limits their potential to be deployed on billions of resource-constrained edge devices. Pruning is a crucial network compression technique, yet most existing methods focus on classification models, with limited attention to detection. Even among those addressing detection, there is a lack of utilization of essential localization information. Also, many pruning methods passively rely on pre-trained models, in which useful and useless components are intertwined, making it difficult to remove the latter without harming the former at the neuron/filter level. To address the above issues, in this paper, we propose a proactive detection-discriminants-based network compression approach for deep visual detectors, which alternates between two steps: (1) maximizing and compressing detection-related discriminants and aligning them with a subset of neurons/filters immediately before the detection head, and (2) tracing the detection-related discriminating power across the layers and discarding features of lower importance. Object location information is exploited in both steps. Extensive experiments, employing four advanced detection models and four state-of-the-art competing methods on the KITTI and COCO datasets, highlight the superiority of our approach. Remarkably, our compressed models can even beat the original base models with a substantial reduction in complexity.
title Visual Detector Compression via Location-Aware Discriminant Analysis
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.17968