# Transform-Invariant Convolutional Neural Networks for Image Classification and Search

arXiv: https://arxiv.org/abs/1912.01447 · HTML: https://mlsanity.com/paper/1912.01447/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Nov 2019.

## Built on by (most cited first, 6 of 6)

- [Generating Unrestricted Adversarial Examples via Three Parameters](https://mlsanity.com/paper/2103.07640/) (2021)
- [Towards Understanding the Effectiveness of Attention Mechanism](https://mlsanity.com/paper/2106.15067/) (2021)
- [Rotated Digit Recognition by Variational Autoencoders with Fixed Output Distributions](https://mlsanity.com/paper/2206.13388/) (2022)
- [Learning Gaussian Data Augmentation in Feature Space for One-shot Object Detection in Manga](https://mlsanity.com/paper/2410.05935/) (2024)
- [OneDConv: Generalized Convolution For Transform-Invariant Representation](https://mlsanity.com/paper/2201.05781/) (2022)
- [Restore Translation Using Equivariant Neural Networks](https://mlsanity.com/paper/2306.16938/) (2023)

## Abstract (arXiv)

Convolutional neural networks (CNNs) have achieved state-of-the-art results on many visual recognition tasks. However, current CNN models still exhibit a poor ability to be invariant to spatial transformations of images. Intuitively, with sufficient layers and parameters, hierarchical combinations of convolution (matrix multiplication and non-linear activation) and pooling operations should be able to learn a robust mapping from transformed input images to transform-invariant representations. In this paper, we propose randomly transforming (rotation, scale, and translation) feature maps of CNNs during the training stage. This prevents complex dependencies of specific rotation, scale, and translation levels of training images in CNN models. Rather, each convolutional kernel learns to detect a feature that is generally helpful for producing the transform-invariant answer given the combinatorially large variety of transform levels of its input feature maps. In this way, we do not require any extra training supervision or modification to the optimization process and training images. We show that random transformation provides significant improvements of CNNs on many benchmark tasks, including small-scale image recognition, large-scale image recognition, and image retrieval. The code is available at https://github.com/jasonustc/caffe-multigpu/tree/TICNN.