Abstract: Disclosed herein is an ocular drug delivery system, comprising a positively charged chitosan-modified hydrogel, a secretory phospholipase A2 (sPLA2) hydrolysable liposome encapsulating an anti-inflammatory drug and having a negative surface charge, a first thiolated hyaluronic acid (HA-SH) nanoparticle encapsulating a wound-healing drug and having a negative surface charge, and a second HA-SH nanoparticle encapsulating a transforming growth factor-? inhibitor and having a negative surface charge. The sPLA2 hydrolysable liposome and the first and second HA-SH nanoparticles are attached to the positively charged chitosan-modified hydrogel by electrostatic attraction. Uses of the system are also disclosed.

Abstract: Disclosed herein is an ocular drug delivery system, comprising a positively charged chitosan-modified hydrogel, a secretory phospholipase A2 (sPLA2) hydrolysable liposome encapsulating an anti-inflammatory drug and having a negative surface charge, a first thiolated hyaluronic acid (HA-SH) nanoparticle encapsulating a wound-healing drug and having a negative surface charge, and a second HA-SH nanoparticle encapsulating a transforming growth factor-? inhibitor and having a negative surface charge. The sPLA2 hydrolysable liposome and the first and second HA-SH nanoparticles are attached to the positively charged chitosan-modified hydrogel by electrostatic attraction. Uses of the system are also disclosed.

Abstract: An apparatus and method can effectively detect both hands and hand shape of a user from images input through cameras. A skin image detecting skin regions from one of the input images and a stereoscopic distance image are used. For hand detection, background and noise are eliminated from a combined image of the skin image and the distance image and regions corresponding to actual both hands are detected from effective images having a high probability of hands. For hand shape detection, a non-skin region is eliminated from the skin image based on the stereoscopic distance information, hand shape candidate regions are detected from the remaining region after elimination, and finally a hand shape is determined.

Abstract: The present invent ion provides a condensed SVM for high-speed learning using a large amount of training data. A first stage WS selector samples a plurality of training data from a training data DB, selects an optimal training vector xt among the plurality of training data, and outputs it to the WS manager. After the first stage finishes, a second stage WS selector extracts training data one by one from the training data DB and selects training data xt satisfying optimality and outputs it to the WS manager. An SVM optimizer extracts training data closest to the training data xt selected by the first and second stage WS selectors from the WS being managed by the WS manager, and condenses the two first and second training data to one training data when the distance between these is smaller than a predetermined value.

Abstract: An SV reduction method for multi-class SVMs is provided with which a number of SVs contained in the multi-class SVMs can be reduced without becoming trapped in a local minimum optimization solution and the reduction of the SVs can be performed at high precision and high speed. The method includes a step of selecting, from a plurality of initially present support vectors, support vector pairs zi, zj (i, j=1, 2, . . . , NS); a step of preparing a single-variable objective function with a single global maximum and determining a maximum value k of the objective function; and a step of applying the maximum value k to the support vector pairs zi and zj to determine a temporary vector Ztemp[i] of small classification errors; and the support vector pairs zi, zj are represented by the temporary vector Ztemp[i].

Abstract: An image processing apparatus and method for real time motion detection is provided. In the apparatus, a sub-sampling module receives and sub-samples a current image and a plurality of previous images and a census transform module performs census transform on each of the sub-sampled images to obtain a census vector. A correlation calculation module calculates and compares correlation values between the current image and the plurality of previous images and detects a region having highest correlation. A motion detection module tracks positions of pixels corresponding to the region having the highest correlation to detect motion information in the images. The image processing apparatus and method can obtain, in real time, the direction and speed of an object that is in motion in each image.

Abstract: An apparatus and method can effectively detect both hands and hand shape of a user from images input through cameras. A skin image detecting skin regions from one of the input images and a stereoscopic distance image are used. For hand detection, background and noise are eliminated from a combined image of the skin image and the distance image and regions corresponding to actual both hands are detected from effective images having a high probability of hands. For hand shape detection, a non-skin region is eliminated from the skin image based on the stereoscopic distance information, hand shape candidate regions are detected from the remaining region after elimination, and finally a hand shape is determined.

Abstract: The present invent ion provides a condensed SVM for high-speed learning using a large amount of training data. A first stage WS selector samples a plurality of training data from a training data DB, selects an optimal training vector xt among the plurality of training data, and outputs it to the WS manager. After the first stage finishes, a second stage WS selector extracts training data one by one from the training data DB and selects training data xt satisfying optimality and outputs it to the WS manager. An SVM optimizer extracts training data closest to the training data xt selected by the first and second stage WS selectors from the WS being managed by the WS manager, and condenses the two first and second training data to one training data when the distance between these is smaller than a predetermined value.

Abstract: An SV reduction method for multi-class SVMs is provided with which a number of SVs contained in the multi-class SVMs can be reduced without becoming trapped in a local minimum optimization solution and the reduction of the SVs can be performed at high precision and high speed. The method includes a step of selecting, from a plurality of initially present support vectors, support vector pairs zi, zj (i, j=1, 2, . . . , NS); a step of preparing a single-variable objective function with a single global maximum and determining a maximum value k of the objective function; and a step of applying the maximum value k to the support vector pairs zi and zj to determine a temporary vector Ztemp[i] of small classification errors; and the support vector pairs zi, zj are represented by the temporary vector Ztemp[i].

Abstract: An image processing apparatus and method for real time motion detection is provided. In the apparatus, a sub-sampling module receives and sub-samples a current image and a plurality of previous images and a census transform module performs census transform on each of the sub-sampled images to obtain a census vector. A correlation calculation module calculates and compares correlation values between the current image and the plurality of previous images and detects a region having highest correlation. A motion detection module tracks positions of pixels corresponding to the region having the highest correlation to detect motion information in the images. The image processing apparatus and method can obtain, in real time, the direction and speed of an object that is in motion in each image.