Object Classification by Learning Their Colors

Step 1. Run Program6Z_1.exe

Step 2. Set the convergence rate (Button ①) to 0.0001 and use the button ② to set the number of repetitions to 100.

Step 3. Click the ‘Learn’ (Button ③) and wait for the 100 iterations completed. Then click button ④ to check the training result from the image shown below.

Step 4. Click one of the four figures located above the “Click one of the above” (Button ①). In the image above the second picture (summer) is clicked.

Step 5. Click ‘Compute Result’ button (②) and check that the small image of the same picture (Summer) is displayed near ①. This is the output from the neural network when the Summer picture is input to the network.

Step 6. Finally, click the ‘Classify’ button (③) to see which season the input image belongs to. Note that there are four numbers displayed in the bottom row, one below each of the four seasons. They represent distances from the input image to the average RGB colors for the four seasons.

KamelConsulting, LLC.

CEO. Byung S. Moon | Tel. +82-10-6627-4306 | ㅣ Biz License 314-86-26086

2F, DICIA ,20 Daedeok-daero 512, Daejeon, S. Korea

2022 Kamel Consulting, LLC