Human-centered deep compositional model for handling occlusions

Gregor Koporec (Author), Janez Perš (Author)

Abstract

Despite their powerful discriminative abilities, Convolutional Neural Networks (CNNs) lack the properties of generative models. This leads to a decreased performance in environments where objects are poorly visible. Solving such a problem by adding more training samples can quickly lead to a combinatorial explosion, therefore the underlying architecture has to be changed instead. This work proposes a Human-Centered Deep Compositional model (HCDC) that combines low-level visual discrimination of a CNN and the high-level reasoning of a Hierarchical Compositional model (HCM). Defined as a transparent model, it can be optimized to real-world environments by adding compactly encoded domain knowledge from human studies and physical laws. The new FridgeNetv2 dataset and a mixture of publicly available datasets are used as a benchmark. The experimental results show the proposed model is explainable, has higher discriminative and generative power, and better handles the occlusion than the current state-of-the-art Mask-RCNN in instance segmentation tasks.

Keywords

računalniški vid;globoko učenje;konvolucijske nevronske mreže;hierarhični kompozicionalni model;zakrivanje;interpretabilnost;poznavanje področja;computer vision;deep learning;convolutional neural networks;hierarchical compositonal model;occlusion;discriminability;generalizability;interpretability;domain knowledge;

Data

Language:	English
Year of publishing:	2023
Typology:	1.01 - Original Scientific Article
Organization:	UL FE - Faculty of Electrical Engineering
UDC:	004
COBISS:	142438403
ISSN:	0031-3203
Views:	18
Downloads:	5
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	Slovenian
Secondary keywords:	računalniški vid;globoko učenje;konvolucijske nevronske mreže;hierarhični kompozicionalni model;zakrivanje;interpretabilnost;poznavanje področja;
Type (COBISS):	Article
Pages:	str. 1-14
Issue:	ǂVol. ǂ138, [article no.] ǂ109397
Chronology:	2023
DOI:	10.1016/j.patcog.2023.109397
ID:	19861989