To Design an Interactive Learning System for Child

To Design an Interactive Learning System for Child by Integrating Blocks with Kinect Tamkang University Taiwan Presenter : FENG-CHIH HSU 1

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 2

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 3

Introduction Building blocks p Develop children’s independent thinking p Use children’s hands and fingers to grasp objects p Cultivate children’s hand-eye coordination 4

Introduction Life of George toys https: //www. youtube. com/watch? v=1 DHZw. SOVKBY 5

Introduction Life of George toys p These blocks and smart phone windows is smaller for children p The Identification can be used to 2 D Identification only p The Identification is affected easily by light and shadow 6

Introduction Interactive block-building system integrate the p We use bigger Lego blocks , PC windows, children using easy p Depth information of kinect can be used to 3 D Identification p Depth information is not affected by light and shadow 7

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 8

The Interactive Block-Building System Step 1. Extracting Depth Information p Our system employs the Kinect to extract the depth information p PC convert the depth information to point cloud dataset with 3 D format 9

The Interactive Block-Building System Step 2. Segmenting the Test Object p We apply algorithm of Random Sample Consensus (RANSAC). p The algorithm remove the point could data of the desk. p Only the point could dataset of assembled object is retained for recognition. 10

The Interactive Block-Building System Step 3. Feature Extraction and Object Recognition p We use the density projection algorithm, the point cloud dataset is projected to the XY, XZ, YZ planes, respectively. 11

The Interactive Block-Building System Step 3. Feature Extraction and Object Recognition p Each plane is then uniformly divided into 16× 12 image blocks. p Each image block is counted as the feature to represent the object whose ranges from 0 to 255. 12

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 13

Experimental Results Ten types of assembled objects are used in the system Total of 460 and 230 images acquired by the Kinect are used as training and testing patterns The recognition accuracy is 95% 14

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 15

e-Block System http: //youtu. be/kg. CPUjf 4 Yq. Q 16

Outline Introduction The Interactive Block-Building System Experimental Results e-Block System Conclusions 17

Conclusions The system help Boosts the motivation of children to assemble blocks. p Train children’s differentiate shapes. p Develop children’s independent thinking. p 18

19

20

21

22

23

Thanks for your attention 24