Time-lapse images are taken to allow embryologists to track how well an embryo is developing, but manual analysis of these images is time-consuming. AI tools have been developed that analyse these images to classify embryos as good or poor quality, but these tools do not work well with the poor quality of many time-lapse images. Time-lapse imaging, whereby regular images are taken of the embryo, is used to improve assessment by providing the embryologist with more information, however analysing this information is time consuming and often involves analysing multiple images of an embryo taken at the same time. To tackle this challenge researchers at Kaunas University of Technology decided to automate the fusion of time-lapse images taken of embryos, in order to create a better-quality image for analysis by embryologists. The resulting fused images were clearer than the individual images and the two embryologists who took part in the study found they were up to three times faster analysing the fused images than the separate images.

The automatic algorithms for perform at least as good as the average embryologist for blastocyst grading and indirectly for predicting fetal heart beat as described above. Developing the algorithm based on time-lapse sequences lead to an improved accuracy compared to using only still images. Training of deep learning algorithms is only based on raw image sequences and requires no prior knowledge of embryology. Thus, the algorithm learns by itself to extract the temporal and the morphological features that are most important for prediction of blastocyst grading. It is important to note that in order to design and train a deep neural network, a substantial amount of data (in this case image sequences) is required.