Repeated Pattern Detection Using CNN Activations

Abstract

We propose a new approach for detecting repeated patterns on a grid in a single image. To do so, we detect repetitions in the space of pre-trained deep CNN filter responses at all layer levels. These encode features at several conceptual levels (from low-level patches to high-level semantics) as well as scales (from local to global). As a result, our repeated pattern detector is robust to challenging cases where repeated tiles show strong variation in visual appearance due to occlusions, lighting or background clutter. Our method contrasts with previous approaches that rely on keypoint extraction, description and clustering or on patch correlation. These generally only detect low-level feature clusters that do not handle variations in visual appearance of the patterns very well. Our method is simpler, yet incorporates high level features implicitly. As such, we can demonstrate detections of repetitions with strong appearance variations, organized on a nearly-regular axis-aligned grid Results show robustness and consistency throughout a varied database of more than 150 images.

3D Façade Labeling over Complex Scenarios: A Case Study Using Convolutional Neural Network and Structure-From-Motion

“…Therefore, certain constraints could be imposed on the relation, for example between a pedestrian and street or vehicles in applications involving self-driving [29,30], such that distance and speed constraints could be imposed between these detected objects. Lettry, L., et al used CNN to detect repeating features in rectified façade images, wherein the repeated patters were verified on a projected grid [81]. Then, it was used as a device to detect those regular characteristics and reconstruct the scene.…”

Robust image retrieval for lacy and embroidered fabric

“…21 Lin et al 22 introduced an effective supervised convolutional neural network (CNN) framework that can simultaneously learn image representations and binary codes for rapid image retrieval. Lettry et al 23 explored the capabilities of pre-trained CNN to detect repetitions in images. However, the above methods based on deep learning are under the assumption that the data are labeled.…”

Моделювання Пошуку Невідомих Локально-Повторюваних Патернів На Зображеннях

“…Підхід на основі стиснення достатньо ефективним робить знаходження точних повторень образів у змісті зображення [1]. На відміну від підходу з використанням тренованої згорткової нейронної мережі [2], метод для знаходження невідомих повторюваних патернів [1] є найбільш вдалим варіантом для розрахунку та класифікації отриманих результатів. У роботах [1,3] використовується контекстне моделювання обмеженого порядку для розпізнавання змісту зображення та виявлення повторюваних патернів.…”