Uncategorized

In this article, we discuss the performance of a sensor system with 12 input features and a time-domain dataset, accompanied by a 4-sensor label output with time-domain values. To evaluate the model’s performance, we focus on the Mean Absolute Error (MAE) metric. We’ll explore the significance of the MAE values obtained and the implications for… Read More »Understanding and Assessing the Performance of a Sensor System with Mean Absolute Error

In this article, we explore how we utilized ChatGPT, specifically the GPT-4 model, to identify the best artificial neural network library and model for a system with 12 sensor input features and a time-domain dataset, accompanied by a 4-sensor label output with time-domain values. Not only did GPT-4 provide valuable insights, but it also assisted… Read More »Enhancing Real-Time Predictions with GPT-4 for a Time-Domain Sensor System

The screenshots are for only 1 sensor from OpenMuscle12 and only 1 finger from LASK4. Download the full dataset and play around with it and tell us what you think.Here is the raw data for the feature and label data taken from Open Muscle:

Introducing the LASK4 System: The Biometric Prosthetic Sensor Label Acquiring Device (prototype Stage) The LASK4 system is a groundbreaking device designed to aid individuals who require hand rehabilitation due to surgeries or other circumstances. But what exactly is the LASK4 system and how does it work? The LASK4 system, also known as the Label Seeking… Read More »LASK4 Update – The Label Seeking AI Companion to Open Muscle

The LASK4 systems was designed to help me train artificial neural networks with finger movement data. These labels are what we use along side Open Muscle to find the least expensive prosthetic sensor. All released opensource and open hardware, find out what we are up to.

The last Open Muscle version 12 had two variations: The 6X2 and the 12X1 named based on their sensor configuration. The next Open Muscle system will be called Open Muscle 24 or when referenced on github or in code: OM-24. There will be only one variation this time around but we are also designing the… Read More »Open Muscle 24 is coming soon!

Deep learning has shown a great promise in many areas of technology were data widely available. There are certain areas that deep learning can be applied to but the infrastructure for data collection is young and not yet matured. Open Muscle aims to supply not just the hardware for biometric muscle contraction data collection but… Read More »Deep Learning in Prosthetic Sensors

We are excited to start capturing the labeled training data with the aide of the new open muscle lask device which is a modified physical therapy tool aimed at strengthening the fingers and dexterity of the hand. We did a livestream working with the device here: You can also view the whole stream video on… Read More »Open Muscle Lask System

Working on the documentation for open muscle there were a few definitions or naming conventions that we needed to settle on. Since we are primarily approaching this problem from a regression training approach with labels we wanted to give the physical label device a name. LASK seemed to be a good fit and we are… Read More »Naming Convention for Open Muscle