Published on Tue May 08 2018

Learning Short-Cut Connections for Object Counting

Daniel Oñoro-Rubio, Mathias Niepert, Roberto J. López-Sastre

Object counting is an important task in computer vision due to its growing demand. We introduce a novel counting model, named Gated U-Net (GU-Net) GU-Nets consistently outperform the base U- net architecture and achieve state-of-the-art performance.

0
0
0
Abstract

Object counting is an important task in computer vision due to its growing demand in applications such as traffic monitoring or surveillance. In this paper, we consider object counting as a learning problem of a joint feature extraction and pixel-wise object density estimation with Convolutional-Deconvolutional networks. We introduce a novel counting model, named Gated U-Net (GU-Net). Specifically, we propose to enrich the U-Net architecture with the concept of learnable short-cut connections. Standard short-cut connections are connections between layers in deep neural networks which skip at least one intermediate layer. Instead of simply setting short-cut connections, we propose to learn these connections from data. Therefore, our short-cuts can work as gating units, which optimize the flow of information between convolutional and deconvolutional layers in the U-Net architecture. We evaluate the introduced GU-Net architecture on three commonly used benchmark data sets for object counting. GU-Nets consistently outperform the base U-Net architecture, and achieve state-of-the-art performance.

Sat Apr 06 2019
Computer Vision
Towards Locally Consistent Object Counting with Constrained Multi-stage Convolutional Neural Networks
High-density object counting in surveillance scenes is challenging mainly due to the drastic variation of object scales. The prevalence of deep learning has boosted the object counting accuracy on several benchmark datasets. We propose a constrained multi-stage Convolutional Neural Networks to jointly pursue locally consistent density map.
0
0
0
Tue Jan 05 2021
Artificial Intelligence
Understanding the Ability of Deep Neural Networks to Count Connected Components in Images
Humans can count very fast by subitizing, but slow substantially as the number of objects increases. We found DNNs do not have the ability to count connected components. We proposed three ML-learnable characteristics to verify learnable problems for ML models.
0
0
0
Wed Jul 25 2018
Computer Vision
Where are the Blobs: Counting by Localization with Point Supervision
Object counting is an important task in computer vision due to its growing demand. State-of-the-art methods use regression-based optimization. These often perform better than detection-based methods that need to learn the more difficult task of predicting the location, size, and shape of each object.
0
0
0
Wed May 26 2021
Computer Vision
Deep Learning for Weakly-Supervised Object Detection and Object Localization: A Survey
Strongly-Supervised Object Detection (WSOD) and Localization (WSOL) are long-standing and challenging tasks in the CV community. Hundreds of WSOD and WSOL methods and techniques have been proposed in the deep learning era.
5
0
1
Mon May 28 2018
Computer Vision
Global Sum Pooling: A Generalization Trick for Object Counting with Small Datasets of Large Images
Conventional global average pooling (GAP) based models are unreliable due to the patchwise cancellation of true overestimates and underestimates for patchwise inference. GSP allows convolutional networks to learn the counting task as a simple linear mapping problem generalized over the input shape and the number of objects present.
0
0
0
Tue Apr 02 2019
Computer Vision
FCOS: Fully Convolutional One-Stage Object Detection
We propose a fully convolutional one-stage object detector (FCOS) FCOS is anchor box free, as well as proposal free. With the only post-processing non-maximum suppression (NMS), FCOS achieves 44.7% in AP.
0
0
0