Defect detection is a common task in industry production, array defect is an undesirable phenomenon in array substrate production process due to the environment, production conditions and so on. Array defect detection is very important for the quality performance of final product. Compared with other defects, array defects have a more complex background, make the detection logic is more advanced and difficult to judge. In this paper, we propose a convolution neural network (CNN) for array defect detection based on Faster-RCNN architecture. On the basis of VGG16 network, we add cross-connection layer in feature extracted layer to improve the accuracy of small-size defect detection. On the other hand, we use ROIAlign layer instead of ROI pooling layer to offset the position migration caused by downsampling. The experimental results show that our strategies have a good performance in array defect detection. The precious of defect recognition is more than 95%, and the recall rate is more than 86% while the defects are divided into 10 categories and the marking image is about 1400 for each type of defect. At the same time, a defect repair scheme based on generating adversarial network (GAN) were proposed: a) input an array defect image and GAN can be used to generate the image without defect or after defect repair, b) the defect repair template can be obtained by compare the generated result with original image, c) this defect repair template can provide reference for practical repair of array defect. In addition, the GAN can be used to optimize array repair scheme and even take the place of manual array defect repair in the future, which can promote the intelligent repair of array defect.

Feature visualization engineering of convolutional neural network is a basic research project in deep learning. The working principle of network, the image features extracted from network and classification basis of images can be revealed, by visualization the extracted features. In this paper, the de‐convolution and CAM method are used to visualize array defect features extracted by CNN. We find that the low‐layer network of CNN is unselective, and its main function is to separate all objects in the picture from their background, this is why the low‐layer networks’ parameters usually do not been trained in transfer learning. Instead, the selectivity is slowly emerging in high‐layer networks and only the object features related to classification are preserved. At the same time, we find that the classification of array defects is not only related to itself, but also to its background, this can also explain why array defect classification is more difficult. On the other hand, this paper verifies the importance of array defect attributes in defect classification by counting the accuracy of network identification after modified the defect attributes, and a defect recognition disturbance rate index was defined to quantify the dependence of defect classification results on its attributes. Finally, the images edited by defect attributes with high disturbance rate are added to the training set, and the defect recognition accuracy is improved by 3% ~ 5%. This method can be used to judge the importance of each array defect attribute, find out the crucial attributes and pinpoint the completeness of training data in CNN classification. At the same time, it can explain the classification principle of CNN and provide necessary guidance and help for attribute image collection and CNN classification accuracy improvement.

We propose a novel measurement method of field of view (FOV) for near‐eye display types (Augmented Reality and Virtual Reality). This method can make use of the existing laboratory optical testing equipment, without additional components, and utilize the characteristics of light propagating along a straight line. In addition, we propose a method for calculating FOV, that is, the edge luminance is reduced to 50% of the central luminance.

In this paper a new method was established to quantitatively analyzed the effects of VHR on the flicker of ADS LCD for both flicker pattern and L127 pattern, we find that for L127 pattern the flicker is mainly related with VHR and obtain their exact relation.