This is a project based on the PCB defect detection using Convolutional Neural Network done under the guidanc of the Dr. hariharan Muthusamy (Associate Professor NIT uk ). the dataset is used from the kaggle containing the 6 type of the defects. i got accuracy of 99.5% in train the dataset shon in the project . the dataset contains the images file along with the annotation file containg the coordinate for the images . the images are cut using the same coordinates with using python . and after that the model is prepared thorugh the dataset . the model is been test and trained with different optimizer using the cnn model .