| Summary: | 碩士 === 開南大學 === 資訊學院碩士班 === 106 === Along with the development of science and technology and human life in pursuit of facilitating life already contains the number of artificial intelligence, the term "artificial intelligence" first appeared in 1950 and allow the PC to be able to target people thinking and learning. Deep learning is modern artificial intelligence of a new technology that depth learning a variety of frameworks, according to google, use depth learning for Robot Arm to arm self-learning Grab and save the entered a large number of self-learning algorithm accuracy is relatively high. Today's smartphone is becoming increasingly popular, according to statistics, Taiwan Intelligent Mobile phone penetration of about 73.4%, that is almost every 3 persons with 2 persons on a smartphone, will use the ready-to-Internet access as well as the characteristics of the lens function can make ready for image recognition APP, see roadside did not understand what items or animals and plants, you only need to take the mobile phone can recognize that can also be some relevant search that you do not need to turn find a figure in view. This study has developed the image using depth learning identify and engage IOS APP in the system, the immediate need to identify items to take immediate items will determine the results displayed. This study uses three algorithms the first algorithm is the use of multi-sensor, MLP multi perceptron layer and the second is the use of roll convolutional neural network base NERVES, third networks, CNN is VGGNET mimics the model of the system to the iPhone APP, back-end is to use PHP with MYSQL for string to capture the user ends the picture will be deposited into the server port and call has been training good module to recognize them, and the results are written to MYSQL, manually after the end of the rite will be showing results.
|
|---|
