CNN Architectures for Geometric Transformation-Invariant Feature Representation in Computer Vision: A Review

Abstract

One of the main challenges in machine vision relates to the problem of obtaining robust representation of visual features that remain unaffected by geometric transformations. This challenge arises naturally in many practical machine vision tasks. For example, in mobile robot applications like simultaneous localization and mapping (SLAM) and visual tracking, object shapes change depending on their orientation in the 3D world, camera proximity, viewpoint, or perspective. In addition, natural phenomena such as occlusion, deformation, and clutter can cause geometric appearance changes of the underlying objects, leading to geometric transformations of the resulting images. Recently, deep learning techniques have proven very successful in visual recognition tasks but they typically perform poorly with small data or when deployed in environments that deviate from training conditions. While convolutional neural networks (CNNs) have inherent representation power that provides a high degree of invariance to geometric image transformations, they are unable to satisfactorily handle nontrivial transformations. In view of this limitation, several techniques have been devised to extend CNNs to handle these situations. This article reviews some of the most promising approaches to extend CNN architectures to handle nontrivial geometric transformations. Key strengths and weaknesses, as well as the application domains of the various approaches are also highlighted. The review shows that although an adequate model for generalized geometric transformations has not yet been formulated, several techniques exist for solving specific problems. Using these methods, it is possible to develop task-oriented solutions to deal with nontrivial transformations.

A Comparison of Pooling Methods for Convolutional Neural Networks

“…As a result, the SPP technique can generate an output with a fixed length without taking the input's size into account. Moreover, the SPP's testing and training phases allow for adaptation to the input image scales, which strengthens the scale-invariance property and eliminates the overfitting problem in the network [50]. However, instead of going into various pooling functions or incorporating learning, spatial pyramid pooling is primarily designed to deal with images of variable sizes and can result in a more complicated learning procedure, resulting in less efficient output, such as a 16:89 percentage error rate on unaugment CIFAR10 [51].…”

A YOLOv6-Based Improved Fire Detection Approach for Smart City Environments

“…Therefore, if there is simply smoke present during the early fire stage, our model waits until it notices a fire. To improve our model and address the aforementioned problem, we are using large datasets, such as JFT-300M [ 68 , 69 , 70 , 71 , 72 ], which comprises 300 million labeled images.…”

An Overview on Visual SLAM: From Tradition to Semantic

“…Figure 13 shows the typical framework of CNN and RNN. CNN can capture spatial features from the image, which help us accurately identify the object and its relationship with other objects in the image [150]. The characteristic of RNN is that it can process an image or numerical data.…”