In this talk, I will present my research in the field of visual anomaly detection with two different approaches which are applicable on the visual and sensor data. In the case of the visual data, I will describe framework using lossy image reconstruction with the autoencoder followed by the feature extraction and one-class classification, which is being used for both industrial and biological data. In the second case, I will present a simplified approach using spectrogram data, pretrained feature extractor and one-class classifier. I will also talk about the experiments which are being performed using both approaches.