Single image calibration using knowledge distillation approaches

Although recent deep learning-based calibration methods can predict extrinsic and intrinsic camera parameters from a single image, their generalization remains limited by the number and distribution of training data samples. The huge computational and space requirement prevents convolutional neural networks (CNNs) from being implemented in resource-constrained environments. This challenge motivated us to learn a CNN gradually, by training new data while maintaining performance on previously learned data. Our approach builds upon a CNN architecture to automatically estimate camera parameters (focal length, pitch, and roll) using different incremental learning strategies to preserve knowledge when updating the network for new data distributions. Precisely, we adapt four common incremental learning, namely: LwF , iCaRL, LU CIR, and BiC by modifying their loss functions to our regression problem. We evaluate on two datasets containing 299008 indoor and outdoor images. Experiment results were significant and indicated which method was better for the camera calibration estimation.

Texture-aware Multi-resolution Image Inpainting

Recent GAN-based inpainting methods have shown remarkable performance using multi-stage networks and/or contextual attention modules (CAM). However, these models require heavy computational resources and may fail to restore realistic texture details. This is mainly due to their training approaches and loss functions. Furthermore, GANs are hard to train on high-resolution images leading to unstable models and poor performance. Inspired by these observations, we propose a novel multi-resolution generators architecture allowing stable training and increased performance. Specifically, our training schema optimizes the parameters of four successive generators such that higher resolution generators exploit the inpainted images produced by lower resolution generators. To restore fine-grained textures, we present a new LBP-based loss function that minimizes the difference between the generated and ground truth textures. We conduct our experiments on Places2 and CelebHQ datasets, and we report qualitative and quantitative results against the state-of-the-art methods. Results show that the computationally efficient model achieves competitive performance.

Learning to Inpaint by Progressively Growing the Mask Regions

Image inpainting is one of the most challenging tasks in computer vision. Recently, generative-based image inpainting methods have been shown to produce visually plausible images. However, they still have difficulties to generate the correct structures and colors as the masked region grows large. This drawback is due to the training stability issue of the generative models. This work introduces a new curriculum-style training approach in the context of image inpainting. The proposed method increases the masked region size progressively in training time, during test time the user gives variable size and multiple holes at arbitrary locations. Incorporating such an approach in GANs may stabilize the training and provides better color consistencies and captures object continuities. We validate our approach on the MSCOCO and CelebA datasets. We report qualitative and quantitative comparisons of our training approach in different models.