We propose a platform based on neural networks to solve the image-to-image translation problem in the context of squeeze ﬂow of micro-droplets . In the ﬁrst part of this paper, we present the governing partial diﬀerential equations to lay out the underlying physics of the problem. We also discuss our developed Python package, sqﬂow , which can potentially serve as free, ﬂexible, and scalable standardized benchmarks in the ﬁelds of machine learning and computer vision. In the second part of this paper, we introduce a residual convolutional neural network to solve the corresponding inverse problem: to translate a high-resolution (HR) imprint image with a speciﬁc liquid ﬁlm thickness to a low-resolution (LR) droplet pattern image capable of producing the given imprint image for an appropriate spread time of droplets. We propose a neural network architecture that learns to systematically tune the reﬁnement level of its residual convolutional blocks by using the function approximators that are trained to map a given input parameter (ﬁlm thickness) to an appropriate reﬁnement level indicator. We use multiple stacks of convolutional layers the output of which is translated according to the reﬁnement level indicators provided by the directly-connected function approximators. Together with a non-linear activation function, such a translation mechanism enables the HR imprint image to be reﬁned sequentially in multiple steps until the target LR droplet pattern image is revealed. The proposed platform can be potentially applied to data compression and data encryption. The developed package and datasets are publicly available on GitHub at https://github.com/sqflow/sqflow . Data encryption, Image-to-image translation, Multi-class multi-label classiﬁcation

Squeeze flow of micro-droplets: convolutional neural network with trainable and tunable refinement

Super-resolution (SR), the process of obtaining highresolution images from one or more low-resolution observations of the same scene, has been a very popular topic of research in the last few decades in both signal processing and image processing areas. Due to the recent developments in Convolutional Neural Networks [1], the popularity of SR algorithms has skyrocketed as the barrier of entry has been lowered significantly. Recently, this popularity has spread into video processing areas to the lengths of developing SR models that work in real-time. In this paper, we compare different SR models that specialize in single image processing and will take a glance at how they evolved to take on many different objectives and shapes over the years.

Single Image Super-Resolution Methods: A Survey

AbstractAmong the recent developments in bioinformatics and chemogenomics, various deep learning methods have been the most prevalent [4, 5, 9]. This resulted in an over-saturation of powerful models that easily pushed the limits of existing datasets. Subsequently, many novel advancements have been done with improvements to the datasets. Amidst these advancements, researchers of Deep Compound Profiler (DeepCOP) [10] set themselves apart with a novel method of introducing new features whilst keeping the deep learning model relatively basic. In this study, we propose to take this novel method one step further by applying transfer learning between cell lines. In order to better evaluate the benefits of transfer learning, we’ve introduced 2 drug-based data splits. The transfer learning method, as its core, utilizes the learned knowledge of “source” cell lines to give a head start to “target” cell lines. Taking advantage of prior knowledge from source cell lines not only boosts existing compounds’ effect prediction, but it can also be used as a premonition for the compounds’ (explicit to the source), effects on target cell lines before they are tested in real life. Our experiments showed improvements up to 22.81% improvement on area under ROC curve (AUC) on the split closest to a wet lab experiment.KeywordsTransfer learningChemogenomicsDomain adaptation 

Bahattin Can Maral

Papers

Single Image Super-Resolution Methods: A Survey

Transfer Learning for Predicting Gene Regulatory Effects of Chemicals