Vision: Overviews

NeurIPS 2019 Schedule


Medical imaging and radiology are facing a major crisis with an ever-increasing complexity and volume of data along an immense economic pressure. Machine learning has emerged as a key technology for developing novel tools in computer aided diagnosis, therapy and intervention. Still, progress is slow compared to other fields of visual recognition, which is mainly due to the domain complexity and constraints in clinical applications, i.e., robustness, high accuracy and reliability. "Medical Imaging meets NeurIPS" aims to bring researchers together from the medical imaging and machine learning communities to discuss the major challenges in the field and opportunities for research and novel applications. The proposed event will be the continuation of a successful workshop organized in NeurIPS 2017 and 2018 (

OpenCV #004 Common Types of Noise Master Data Science


Highlights: We will give an overview of the most common types of noise that is present in images. We will show how we can generate these types of noise and add them to clean images. Then, we will show how we can filter these images using a simple median filter. In this post, we will assume that we "know" how the noise looks like in our experiments and then it will be easier for us to find an optimal way how to remove that noise. Different kind of imaging systems might give us different noise.

5 Significant Object Detection Challenges and Solutions


Object detection problems pose several unique obstacles beyond what is required for image classification. Five such challenges are reviewed in this post along with researchers' efforts to overcome these complications. The field of computer vision has experienced substantial progress recently owing largely to advances in deep learning, specifically convolutional neural nets (CNNs). Image classification, where a computer classifies or assigns labels to an image based on its content, can often see great results simply by leveraging pre-trained neural nets and fine-tuning the last few throughput layers. Classifying and finding an unknown number of individual objects within an image, however, was considered an extremely difficult problem only a few years ago.

Embodied Agent - an overview


For an autonomous embodied agent acting in the real world (e.g., an animal, a human, or a robot), perceptual categorization--the ability to make distinctions--is a hard problem (Harnad, 2005). First, based on the stimulation impinging on its sensory arrays (sensation) the agent has to rapidly determine and attend to what needs to be categorized. Second, the appearance and properties of objects or events in the environment being classified fluctuate continuously, for example owing to occlusions, or changes of distances and orientations with respect to the agent. And third, the environmental conditions (e.g., illumination, viewpoint, and background noise) vary considerably. There is much relevant work in computer vision that has been devoted to extracting scale- and translation-invariant low-level visual features and high-level multidimensional representations for the purpose of robust perceptual categorization (Riesenhuber & Poggio, 2002).

Recent Advances in Modern Computer Vision


In the last 50 years, computers have learned to count and classify but still weren't able to see until now. Today, as of 2019, the field of computer vision is rapidly flourishing, holding vast potential to alleviate everything from healthcare disparities to mobility limitations on a global scale. In recent years, we have seen great success in Computer Vision built on top of AlexNet or similar CNN based architectures as a backbone. It's true that the process is modeled after the human brain in terms of how it learns; a network of learning units called neurons learn how to convert input signals such as a picture of a house into corresponding output signals like the label'house'. For more details regarding this see my previous blog.

AI Provides a Detailed Road Map for Interventional Lung Procedures


Precise medical imaging and analysis could enable early detection of lung cancer, help determine its exact size and location, and significantly improve diagnosis and treatment. This is usually done in a process called segmentation, which uses computers to identify the boundaries of the lung from surrounding thoracic tissue on CT images. From this process, a detailed 3-D map of the airways may be generated that can help to plan and navigate a bronchoscopy procedure to obtain biopsy samples and to perform other clinical interventions. "Until now, this process was very difficult because you need the radiologist, or even the surgeon, to spend much time to understand how to get to the specific place [where the lesion is located]. And this is sometimes prone to error," said Ron Soferman, founder and CEO of RSIP Vision, in an interview with MD DI. "It's very critical [to know the precise location] because, if you miss the lesion, you will take a biopsy from some random part of the lung and it will give a negative result."

Theory of Spectral Method for Union of Subspaces-Based Random Geometry Graph Machine Learning

Union of Subspaces (UoS) model serves as an important model i n statistical machine learning. Briefly speaking, UoS models those high-dimensional da ta, encountered in many real-world problems, which lie close to low-dimensional subspaces corresponding to several classes to which the data belong, such as handwritten digits (Hasti e and Simard, 1998), face images (Basri and Jacobs, 2003), DNA microarray data (Parvare sh et al., 2008), and hyper-spectral images (Chen et al., 2011), to name just a few. A fund amental task in processing data points in UoS is to cluster these data points, which is kn own as Subspace Clustering (SC). Applications of SC has spanned all over science and eng ineering, including motion segmentation (Costeira and Kanade, 1998; Kanatani, 2001), face recognition (Wright et al., 2008), and classification of diseases (McWilliams and Monta na, 2014) and so on. We refer the reader to the tutorial paper (Vidal, 2011) for a review of the development of SC. The authors are with Department of Electronic Engineering, Tsinghua University, Beijing 100084, China. The corresponding author of this paper is Y. Gu (gyt@tsinghu

A General Framework for Complex Network-Based Image Segmentation Machine Learning

With the recent advances in complex networks theory, graph-based techniques for image segmentation has attracted great attention recently. In order to segment the image into meaningful connected components, this paper proposes an image segmentation general framework using complex networks based community detection algorithms. If we consider regions as communities, using community detection algorithms directly can lead to an over-segmented image. To address this problem, we start by splitting the image into small regions using an initial segmentation. The obtained regions are used for building the complex network. To produce meaningful connected components and detect homogeneous communities, some combinations of color and texture based features are employed in order to quantify the regions similarities. To sum up, the network of regions is constructed adaptively to avoid many small regions in the image, and then, community detection algorithms are applied on the resulting adaptive similarity matrix to obtain the final segmented image. Experiments are conducted on Berkeley Segmentation Dataset and four of the most influential community detection algorithms are tested. Experimental results have shown that the proposed general framework increases the segmentation performances compared to some existing methods.

An Overview of Human Pose Estimation with Deep Learning


A Human Pose Skeleton represents the orientation of a person in a graphical format. Essentially, it is a set of coordinates that can be connected to describe the pose of the person. Each coordinate in the skeleton is known as a part (or a joint, or a keypoint). A valid connection between two parts is known as a pair (or a limb). Note that, not all part combinations give rise to valid pairs.

Detecting Bias with Generative Counterfactual Face Attribute Augmentation Machine Learning

We introduce a simple framework for identifying biases of a smiling attribute classifier. Our method poses counterfactual questions of the form: how would the prediction change if this face characteristic had been different? We leverage recent advances in generative adversarial networks to build a realistic generative model of face images that affords controlled manipulation of specific image characteristics. We introduce a set of metrics that measure the effect of manipulating a specific property of an image on the output of a trained classifier. Empirically, we identify several different factors of variation that affect the predictions of a smiling classifier trained on CelebA.