Biological - AI

Datamining for neural computational mechanism

---
B. Yousefi's first Ph.D. in Datamining for neural computational mechanism, Department of Artificial Intelligence- University of Malaya
---
Structure overall, visual system analytical models. The approach pursues to develop the computational models for recognition of biological movements and characterize the responses for different actions. This model is the perspective of the original model consists of particular computations of slow and fast feature data. The model can operate in wide range of high-dimension of input and the outcome is a combination of the ventral and dorsal processing stream.

Studies on computational neuroscience through functional magnetic resonance imaging and following human visual systems state that the mammalian brain pursues two distinct pathways in the model. These pathways are designed to analyze not only motion information (optical flow) but also the ventral processing stream in the brain that proceeds with form features, in which Gabor wavelet is widely used. The original model of the mammalian visual system represents two independent pathways, which become a subject of interest among researchers. Model development is performed via systematic organization, where the active basis model is added into the ventral processing stream. The Gabor wavelet-based and supervised method is efficient in terms of Gabor beam utilization and object recognition-directed task through form pathway. In addition, the motion information that is generated via optical flow in motion pathway is stabilized through applying the fuzzy membership scoring, which delays the changes in optical flow outcomes and provides further robustness to the system. The interaction between these processing pathways is another substantial matter implied in the model. The cross-connection of the two pathways is implied throughout the present research via direct consideration, such as shared sketch algorithm and optical flow information, fuzzy max-product involvement, and scoring among each other.

The hierarchical model follows the original model and interpretation of the data is in the perspective of combination of slowness and fast features provided from ventral and dorsal processing stream. An overview of these two pathways, form and motion pathways is revealed. Insert depicts the various types of neural detectors in diverse parts of hierarchy. V1 and IT represent primary visual cortex and inferotemporal cortex also KO and STS are kinetic occipital cortex and superior temporal sulcus respectively. These abbreviations along with others indicate visual cortex in monkey and human (Giese & Poggio, 2003).

In addition, the model is considered a form information through active basis model based on incremental slow feature analysis (denoted as slow features). In this study, the motion perception in human visual system comprises fast and slow feature interactions, which render biological movement understandable. Primarily, a form feature is defined. This feature biologically follows the visual system through applying active basis model and incremental slow feature analysis for extraction of the slowest form features of human object for ventral stream. The interaction is considered within the time that provides valuable features to recognize biological movements. Incremental slow feature analysis provides a chance for fast action prototypes and bag-of-word techniques, and opens a new perspective to recognize the original biological movement model. Episodic observation is required to extract the slowest features.

The schematic of model is presented here for both pathways. In ventral processing stream, form pathway, a set of Gabor filters have been applied at different orientations, positions and phases; outputs of V1 part is outcomes of quadrature-phase pairs, summed, squared, and square-rooted. Then outputs of the filter normalize considering local population. Afterwards filter outcomes are max pools and summed across space. The MAX, SUM operations are based on the attained active bases of the object form. This initial part of the schematic are done through using ABM (Si et al., 2010) as Gabor based object recognition operation. Finally, the output of the ABM is utilized into the slowness principle method (incremental slow feature analysis) (Kompella et al., 2011; 2012) for extraction of form slow features. On the other hand in the dorsal processing stream, which helps to obtain motion information throughout the high-dimensional input stream. Motion pathway is attained using Optical Flow. Average of these flows within the episode (t0;t1;...;tn) plays the fast features in this hierarchy which temporally ventral stream requires for utilizing IncSFA for generation slow features. However, each pathways can has its own decision in categorization and it justifies two patients (DF and RV) performances (Goodale et al., 1994).

However, fast features of dorsal processing pathway through episodic ventral analysis update the processing of motion information. Experimental results in the development of the biological movement model indicate promising accuracies for proposed improvements and favorable performance on different datasets (KTH and Weizmann). The results also provide promising direction on this area.