KL-DNAS: Knowledge Distillation-Based Latency Aware-Differentiable Architecture Search for Video Motion Magnification.

Video motion magnification is the task of making subtle minute motions visible. Many times subtle motion occurs while being invisible to the naked eye, e.g., slight deformations in muscles of an athlete, small vibrations in the objects, microexpression, and chest movement while breathing. Magnification of such small motions has resulted in various applications like posture deformities detection, microexpression recognition, and studying the structural properties. State-of-the-art (SOTA) methods have fixed computational complexity, which makes them less suitable for applications requiring different time constraints, e.g., real-time respiratory rate measurement and microexpression classification. To solve this problem, we propose a knowledge distillation-based latency aware-differentiable architecture search (KL-DNAS) method for video motion magnification. To reduce memory requirements and to improve denoising characteristics, we use a teacher network to search the network by parts using knowledge distillation (KD). Furthermore, search among different receptive fields and multifeature connections are applied for individual layers. Also, a novel latency loss is proposed to jointly optimize the target latency constraint and output quality. We are able to find 2.8 × smaller model than the SOTA method and better motion magnification with lesser distortions. https://github.com/jasdeep-singh-007/KL-DNAS.

Adding value to banana farming: Antibody production in post-harvest leaves.

Banana, a globally popular fruit, is widely cultivated in tropical and sub-tropical regions. After fruit harvest, remaining banana plant materials are low-value byproducts, mostly composted or used as fibre or for food packaging. As an aim to potentially increase farmer income, this study explored underutilised banana biomass as a novel plant tissue for production of a high-value product. Protein scFvTG130 used in this study, is an anti-toxoplasma single chain variable fragment antibody that can be used in diagnostics and neutralising the Toxoplasma gondii pathogen. Using detached banana leaves, we investigated the factors influencing the efficacy of a transient expression system using reporter genes and recombinant protein, scFvTG130. Transient expression was optimal at 2 days after detached banana leaves were vacuum infiltrated at 0.08 MPa vacuum pressure for a duration of 3 min with 0.01% (v/v) Tween20 using Agrobacterium strain GV3101 harbouring disarmed virus-based vector pIR-GFPscFvTG130. The highest concentration of anti-toxoplasma scFvTG130 antibody obtained using detached banana leaves was 22.8 µg/g fresh leaf tissue. This first study using detached banana leaf tissue for the transient expression of a recombinant protein, successfully demonstrated anti-toxoplasma scFvTG130 antibody expression, supporting the potential application for other related proteins using an underutilised detached banana leaf tissue.