Abstract. The problem of multi-view/view-invariant recognition remains one of the most fundamental challenges to the progress of the computer vision. In this paper we consider the problem of modeling the combined object-viewpoint manifold. The shape and appearance of an object in a given image is a function of its category, style within category, viewpoint, and several other factors. The visual manifold (in any chosen feature representation space) given all these variability collectively is very hard and even impossible to model. We propose an efficient computational framework that can untangle such a complex manifold, and achieve a model that separates a view-invariant category representation, from categoryinvariant pose representation. We outperform the state of the art in the three widely used multiview dataset, for both category recognition, and pose estimation.

There is a growing demand in automated public safety systems for detecting unauthorized vehicle parking, intrusions, un-intended baggage, etc. Object detection and recognition significantly impact these applications. Object detection and recognition are challenging problems in this context, since the purpose of the surveillance videos is to capture a wide landscape of the scene; resulting in small, low-resolution and occluded images for objects. In this paper, we present an experimental study on geometric and appearance features (∈ R ≈25000 ) for outdoor video surveillance systems. We also studied the classification performance under two dimensionality reduction techniques (i.e. PCA and Entropy-Based feature Selection). As a result, we built an experimental framework for an object classification system for surveillance videos with different configurations.

Visual speech recognition is a challenging problem, due to confusion between visual speech features. The speaker identification problem is usually coupled with speech recognition. Moreover, speaker identification is important to several applications, such as automatic access control, biometrics, authentication, and personal privacy issues. In this paper, we propose a novel approach for lipreading and speaker identification. We propose a new approach for manifold parameterization in a low-dimensional latent space, where each manifold is represented as a point in that space. We initially parameterize each instance manifold using a nonlinear mapping from a unified manifold representation. We then factorize the parameter space using Kernel Partial Least Squares (KPLS) to achieve a low-dimension manifold latent space. We use two-way projections to achieve two manifold latent spaces, one for the speech content and one for the speaker. We apply our approach on two public databases: AVLetters and OuluVS. We show the results for three different settings of lipreading: speaker independent, speaker dependent, and speaker semi-dependent. Our approach outperforms for the speaker semi-dependent setting by at least 15% of the baseline, and competes in the other two settings.

Object recognition and pose estimation are two fundamental problems in the field of computer vision. Recognizing objects and their poses/viewpoints are critical components of ample vision and robotic systems. Multiple viewpoints of an object lie on an intrinsic low-dimensional manifold in the input space (i.e. descriptor space). Different objects captured from the same set of viewpoints have manifolds with a common topology. In this paper we utilize this common topology between object manifolds by learning a low-dimensional latent space which non-linearly maps between a common unified manifold and the object manifold in the input space. Using a supervised embedding approach, the latent space is computed and used to jointly infer the category and pose of objects. We empirically validate our model by using multiple inference approaches and testing on multiple challenging datasets. We compare our results with the state-of-the-art and present our increased category recognition and pose estimation accuracy.