Our Projects
Artificial generation of images with vehicles.
Our model needs as input images, the ones that contains vehicles to remove, but as we have not found this situation in different datasets (it’s so difficult to contain exactly the same image with and without vehicle to remove) we have take the decision of generate our dataset artificially.
As a solution, we have added different vehicles extracted from the BDD100K set in an artificial way to other images again from the same set. There will also obtain the image that in our model we will use as input, which is the one with the black mask applied on top.
Below we can see an example of this modification.
The problem of Low-Light Image Enhancement involves improving the lighting levels in dark images. To solve this problem, it is possible to apply convolutional neural networks (CNNs) with supervised learning, meaning that a corrected reference image (“ground truth”) is needed. Along with this image and the image obtained by the model, the loss is calculated, which will be minimized and fed into the ‘backpropagation’ algorithm. In addition to methods that use supervised learning, there are others that use unsupervised learning, meaning that in this case, no reference image is required, and therefore the loss calculation is done differently.
During the image capture process, there are various parameters that define the final result obtained, such as, exposure time, sensor ISO, chosen aperture, or lens focal length. Among these, the ones that are most relevant for nighttime photography are, without a doubt, exposure time, sensor ISO, and lens aperture.
Visibility is a complex issue that detection models struggle with. While on clear days these algorithms works perfectly, they cannot detect objects with the same accuracy under adverse conditions such as rain, haze or snow. On this project we focused on creating a dataset to train these models and achieve a good performance of vehicle detection under adverse weather conditions.
The impact of Artificial Intelligence (AI) in the field of autonomous driving has been substantial in recent years. As a result, there are numerous applications of these techniques to related problems, including gesture recognition, which we will discuss in this post. In addition to this task, many others are also based on computer vision in this field, such as the generation of synthetic scenarios used for training and validating detection or trajectory prediction models.