An Integrated Approach for Video Captioning and Applications

Physical computing infrastructure, data gathering, and algorithms have recently had significant advances to extract information from images and videos. The growth has been especially outstanding in image captioning and video captioning. However, most of the advancements in video captioning still take place in short videos. In this research, we caption longer videos only by using the keyframes, which are a small subset of the total video frames. Instead of processing thousands of frames, only a few frames are processed depending on the number of keyframes. There is a trade-off between the computation of many frames and the speed of the captioning process. The approach in this research is to allow the user to specify the trade-off between execution time and accuracy. In addition, we argue that linking images, videos, and natural language offers many practical benefits and immediate practical applications. From the modeling perspective, instead of designing and staging explicit algorithms to process videos and generate captions in complex processing pipelines, our contribution lies in designing hybrid deep learning architectures to apply in long videos by captioning video keyframes. We consider the technology and the methodology that we have developed as steps toward the applications discussed in this research.

Texture image classification based on a pseudo-parabolic diffusion model

This work proposes a novel method based on a pseudo-parabolic diffusion process to be employed for texture recognition. The proposed operator is applied over a range of time scales giving rise to a family of images transformed by nonlinear filters. Therefore each of those images are encoded by a local descriptor (we use local binary patterns for that purpose) and they are summarized by a simple histogram, yielding in this way the image feature vector. The proposed approach is tested on the classification of well established benchmark texture databases and on a practical task of plant species recognition. In both cases, it is compared with several state-of-the-art methodologies employed for texture recognition. Our proposal outperforms those methods in terms of classification accuracy, confirming its competitiveness. The good performance can be justified to a large extent by the ability of the pseudo-parabolic operator to smooth possibly noisy details inside homogeneous regions of the image at the same time that it preserves discontinuities that convey critical information for the object description. Such results also confirm that model-based approaches like the proposed one can still be competitive with the omnipresent learning-based approaches, especially when the user does not have access to a powerful computational structure and a large amount of labeled data for training.

Learning an Adaptive Model for Extreme Low-light Raw Image Processing

Low-light images suffer from severe noise and low illumination. Current deep learning models that are trained with real-world images have excellent noise reduction, but a ratio parameter must be chosen manually to complete the enhancement pipeline. In this work, we propose an adaptive low-light raw image enhancement network to avoid parameter-handcrafting and to improve image quality. The proposed method can be divided into two sub-models: Brightness Prediction (BP) and Exposure Shifting (ES). The former is designed to control the brightness of the resulting image by estimating a guideline exposure time $t_1$. The latter learns to approximate an exposure-shifting operator $ES$, converting a low-light image with real exposure time $t_0$ to a noise-free image with guideline exposure time $t_1$. Additionally, structural similarity (SSIM) loss and Image Enhancement Vector (IEV) are introduced to promote image quality, and a new Campus Image Dataset (CID) is proposed to overcome the limitations of the existing datasets and to supervise the training of the proposed model. Using the proposed model, we can achieve high-quality low-light image enhancement from a single raw image. In quantitative tests, it is shown that the proposed method has the lowest Noise Level Estimation (NLE) score compared with the state-of-the-art low-light algorithms, suggesting a superior denoising performance. Furthermore, those tests illustrate that the proposed method is able to adaptively control the global image brightness according to the content of the image scene. Lastly, the potential application in video processing is briefly discussed.

Deep Learning based Multi-modal Computing with Feature Disentanglement for MRI Image Synthesis

Purpose: Different Magnetic resonance imaging (MRI) modalities of the same anatomical structure are required to present different pathological information from the physical level for diagnostic needs. However, it is often difficult to obtain full-sequence MRI images of patients owing to limitations such as time consumption and high cost. The purpose of this work is to develop an algorithm for target MRI sequences prediction with high accuracy, and provide more information for clinical diagnosis. Methods: We propose a deep learning based multi-modal computing model for MRI synthesis with feature disentanglement strategy. To take full advantage of the complementary information provided by different modalities, multi-modal MRI sequences are utilized as input. Notably, the proposed approach decomposes each input modality into modality-invariant space with shared information and modality-specific space with specific information, so that features are extracted separately to effectively process the input data. Subsequently, both of them are fused through the adaptive instance normalization (AdaIN) layer in the decoder. In addition, to address the lack of specific information of the target modality in the test phase, a local adaptive fusion (LAF) module is adopted to generate a modality-like pseudo-target with specific information similar to the ground truth. Results: To evaluate the synthesis performance, we verify our method on the BRATS2015 dataset of 164 subjects. The experimental results demonstrate our approach significantly outperforms the benchmark method and other state-of-the-art medical image synthesis methods in both quantitative and qualitative measures. Compared with the pix2pixGANs method, the PSNR improves from 23.68 to 24.8. Conclusion: The proposed method could be effective in prediction of target MRI sequences, and useful for clinical diagnosis and treatment.

Desmoking laparoscopy surgery images using an image-to-image translation guided by an embedded dark channel

In laparoscopic surgery, the visibility in the image can be severely degraded by the smoke caused by the $CO_2$ injection, and dissection tools, thus reducing the visibility of organs and tissues. This lack of visibility increases the surgery time and even the probability of mistakes conducted by the surgeon, then producing negative consequences on the patient's health. In this paper, a novel computational approach to remove the smoke effects is introduced. The proposed method is based on an image-to-image conditional generative adversarial network in which a dark channel is used as an embedded guide mask. Obtained experimental results are evaluated and compared quantitatively with other desmoking and dehazing state-of-art methods using the metrics of the Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM) index. Based on these metrics, it is found that the proposed method has improved performance compared to the state-of-the-art. Moreover, the processing time required by our method is 92 frames per second, and thus, it can be applied in a real-time medical system trough an embedded device.

Towards Learning a Self-inverse Network for Bidirectional Image-to-image Translation

The one-to-one mapping is necessary for many bidirectional image-to-image translation applications, such as MRI image synthesis as MRI images are unique to the patient. State-of-the-art approaches for image synthesis from domain X to domain Y learn a convolutional neural network that meticulously maps between the domains. A different network is typically implemented to map along the opposite direction, from Y to X. In this paper, we explore the possibility of only wielding one network for bi-directional image synthesis. In other words, such an autonomous learning network implements a self-inverse function. A self-inverse network shares several distinct advantages: only one network instead of two, better generalization and more restricted parameter space. Most importantly, a self-inverse function guarantees a one-to-one mapping, a property that cannot be guaranteed by earlier approaches that are not self-inverse. The experiments on three datasets show that, compared with the baseline approaches that use two separate models for the image synthesis along two directions, our self-inverse network achieves better synthesis results in terms of standard metrics. Finally, our sensitivity analysis confirms the feasibility of learning a self-inverse function for the bidirectional image translation.

Towards Zero-shot Cross-lingual Image Retrieval

There has been a recent spike in interest in multi-modal Language and Vision problems. On the language side, most of these models primarily focus on English since most multi-modal datasets are monolingual. We try to bridge this gap with a zero-shot approach for learning multi-modal representations using cross-lingual pre-training on the text side. We present a simple yet practical approach for building a cross-lingual image retrieval model which trains on a monolingual training dataset but can be used in a zero-shot cross-lingual fashion during inference. We also introduce a new objective function which tightens the text embedding clusters by pushing dissimilar texts from each other. Finally, we introduce a new 1K multi-lingual MSCOCO2014 caption test dataset (XTD10) in 7 languages that we collected using a crowdsourcing platform. We use this as the test set for evaluating zero-shot model performance across languages. XTD10 dataset is made publicly available here: https://github.com/adobe-research/Cross-lingual-Test-Dataset-XTD10

Image Transformation Network for Privacy-Preserving Deep Neural Networks and Its Security Evaluation

We propose a transformation network for generating visually-protected images for privacy-preserving DNNs. The proposed transformation network is trained by using a plain image dataset so that plain images are transformed into visually protected ones. Conventional perceptual encryption methods have a weak visual-protection performance and some accuracy degradation in image classification. In contrast, the proposed network enables us not only to strongly protect visual information but also to maintain the image classification accuracy that using plain images achieves. In an image classification experiment, the proposed network is demonstrated to strongly protect visual information on plain images without any performance degradation under the use of CIFAR datasets. In addition, it is shown that the visually protected images are robust against a DNN-based attack, called inverse transformation network attack (ITN-Attack) in an experiment.

Enhance Images as You Like with Unpaired Learning

Low-light image enhancement exhibits an ill-posed nature, as a given image may have many enhanced versions, yet recent studies focus on building a deterministic mapping from input to an enhanced version. In contrast, we propose a lightweight one-path conditional generative adversarial network (cGAN) to learn a one-to-many relation from low-light to normal-light image space, given only sets of low- and normal-light training images without any correspondence. By formulating this ill-posed problem as a modulation code learning task, our network learns to generate a collection of enhanced images from a given input conditioned on various reference images. Therefore our inference model easily adapts to various user preferences, provided with a few favorable photos from each user. Our model achieves competitive visual and quantitative results on par with fully supervised methods on both noisy and clean datasets, while being 6 to 10 times lighter than state-of-the-art generative adversarial networks (GANs) approaches.

Exploring Self-attention for Image Recognition

Recent work has shown that self-attention can serve as a basic building block for image recognition models. We explore variations of self-attention and assess their effectiveness for image recognition. We consider two forms of self-attention. One is pairwise self-attention, which generalizes standard dot-product attention and is fundamentally a set operator. The other is patchwise self-attention, which is strictly more powerful than convolution. Our pairwise self-attention networks match or outperform their convolutional counterparts, and the patchwise models substantially outperform the convolutional baselines. We also conduct experiments that probe the robustness of learned representations and conclude that self-attention networks may have significant benefits in terms of robustness and generalization.