DeePM: A Deep Part-Based Model for Object Detection and Semantic Part Localization

TL;DR

DeePM is a deep, part-based model that improves object detection and semantic part localization by explicitly modeling object-part configurations, outperforming existing R-CNN variants on PASCAL VOC 2012.

Contribution

The paper introduces DeePM, a novel deep graphical model that explicitly represents object-part configurations with flexible sharing, enhancing detection accuracy.

Findings

DeePM outperforms OP R-CNN in object and part detection.

DeePM surpasses Fast and Faster R-CNN in object detection.

Annotated semantic parts for all 20 PASCAL VOC 2012 categories.

Abstract

In this paper, we propose a deep part-based model (DeePM) for symbiotic object detection and semantic part localization. For this purpose, we annotate semantic parts for all 20 object categories on the PASCAL VOC 2012 dataset, which provides information on object pose, occlusion, viewpoint and functionality. DeePM is a latent graphical model based on the state-of-the-art R-CNN framework, which learns an explicit representation of the object-part configuration with flexible type sharing (e.g., a sideview horse head can be shared by a fully-visible sideview horse and a highly truncated sideview horse with head and neck only). For comparison, we also present an end-to-end Object-Part (OP) R-CNN which learns an implicit feature representation for jointly mapping an image ROI to the object and part bounding boxes. We evaluate the proposed methods for both the object and part detection performance on PASCAL VOC 2012, and show that DeePM consistently outperforms OP R-CNN in detecting objects and parts. In addition, it obtains superior performance to Fast and Faster R-CNNs in object detection.