Meeting Room Dashboard

Introduction

The project displays previous, current and upcoming meeting information on the dashboard by fetching data from the outlook calendar. In the outlook calendar, there are meeting room calendars for three different floors. Meeting information about each floor is available in their corresponding floor calendars with raspberry and monitor.

Documentation

https://javrasoftware-my.sharepoint.com/:w:/r/personal/bishal_dangal_javra_com/_layouts/15/Doc.aspx?sourcedoc=%7B2C2CD74E-0132-4104-A56D-F4E528DBEADC%7D&file=Meeting%20Room%20Dashboard.docx&action=default&mobileredirect=true

POC

https://meeting.javra.com/

GIT

https://git.javra.com/internal-rnd/meeting-room-dashboard

Smart Javra Chatbot

javrachatbot

Introduction

Javra’s Smart AI Chatbot, a cutting-edge solution designed to transform your workflow and enhance the overall employee experience. This intelligent companion provides instant access to crucial information, delivering prompt responses that streamline daily tasks and foster collaboration across all levels within your company.

Overview

https://javra.com/News/Smart%20AI%20Chatbot

 

POC

https://smartuat.javra.com/home

 

CV Screening Tool​

 

Introduction

  • tool leverages machine learning models to classify resumes into four distinct categories: Waiting, Qualified, Overqualified, and Not Qualified. 
  • job descriptions are parsed into JSON format using the Gemini API.

Documentation
https://git.javra.com/internal-rnd/resume-screener/-/blob/main/docs/Overview.pdf

POC

https://resume.javra.com/

GIT

https://git.javra.com/internal-rnd/cv_screening

Microsoft Hololens

Introduction

Microsoft Hololens for interactive QR code scanning. By leveraging the advanced capabilities of Hololens, users can scan QR codes in their environment, instantly accessing and interacting with digital content overlaid onto the physical world. This integration enhances user experience, streamlines processes, and opens up new possibilities for applications in various industries, including retail, logistics, and marketing.

Documentation

Holo-lens

Overview
https://javra.com/Innovate-with-HoloLens

 

Vedio

 

Git

https://git.javra.com/internal-rnd/qr-tracking –> Unity part

https://git.javra.com/internal-rnd/qrtrackingservice_vs –> Visual Studio Part

 

Warehouse Path Finder

Introduction

DecaWave 3000 Module is a type of wireless communication module developed by DecaWave. It is based on ultra-wideband (UWB) technology, which enables precise indoor positioning and localization capabilities. The module is designed to provide accurate location data, typically within centimeter-level accuracy, making it suitable for applications such as asset tracking, indoor navigation, industrial automation, and real-time location systems (RTLS).

Overview
https://javra.com/warehouse-path-finder

POC

https://warehouse.javra.com/

Documentation

Route-finder-warehouse

GIT
https://git.javra.com/rnd/ips-poc-decawave

Mood Scene Generator

mood-scenemood-scence

Introduction

The Product Visualization Tool revolutionizes visual realism with stable diffusion algorithms, creating lifelike scenes and authentic representations for diverse applications.

  • Generate mood scenes of product based on prompts
  • Using Stable Diffusion Models

Overview

https://javra.com/Javra-Product-Visualization-Tool

POC

  1. https://background-generate.javra.com/ (for simple text-to-image generator) 
  1. https://background-generate.javra.com/bg_replace.html (for mood scene generator) 

 

Documentation

https://javrasoftware.sharepoint.com/:w:/r/sites/RDinAIML/_layouts/15/Doc2.aspx?action=edit&sourcedoc=%7B6b09bb8a-e028-4e68-9213-994efbfaa2f0%7D&wdOrigin=TEAMS-MAGLEV.teamsSdk_ns.rwc&wdExp=TEAMS-TREATMENT&wdhostclicktime=1722938283631&web=1

GIT

https://git.javra.com/internal-rnd/sd-api

Tomato Trail App

tomato-count

Introduction

  • Identify tomatoes, their count using YoloV9 neural network
  • common colors and size

 

Download

EnzaCounts.EnzaCounts-Signed 8-1 1.apk

 

Documentation

https://javrasoftware-my.sharepoint.com/shared?id=%2Fsites%2FRDinAIML%2FShared%20Documents%2FGeneral%2FTomato%20Trail%20APP&listurl=https%3A%2F%2Fjavrasoftware%2Esharepoint%2Ecom%2Fsites%2FRDinAIML%2FShared%20Documents&viewid=c0a15e4c%2D4fdc%2D43a8%2Da33f%2Dc7630089b33a

GIT

https://git.javra.com/sutha022/enzacounts/-/tree/production?ref_type=heads

 

Clod and Tuber Detection​

ClodnTuber

Introduction

We have been working on a new dataset and trained full-sized images, which are around 2448×2048 pixels, using the latest YOLOv9 model. Due to the larger image size, we have achieved an inference speed of around 55ms. If we optimize it using TensorRT as we did previously, we can reduce the inference time to around 25ms. Furthermore, this new model addresses missed detections of tubes and clods and has improved accuracy. For now, the dataset is sufficient for similar scenarios. We request you to test this newer model. The previous inference speed was on cropped images with a size of around 640 pixels.

We have attached the tuberclod.zip, which contains the detect.py script slightly changed, and the model. You can use:

yolo export model=/path/to/Object_detection/image_objectDetection_yolov9/tuber_clodV9.pt format=engine device=0 half=True

Performance is fine for our goal now: process 1 picture per second.
Accuracy is also good. We executed a test on 100 objects, 50% clod 50% tuber randomly picked. Results:
Test 1: 100 objects 50/50. Seprately run in sequence: 3 tubers false positive as clod. No false results in tubers. 5 objects not ejected after first run (this is mechanical, not in software).
Test 2: 100 objects 50/50. all objects mixed: 2 tubers false positive as clod . No false results in tubers. 6 not ejected (mechanical issue).
so, if we pick te full population, we see a 100% score on tubers, 97% on clods.
Documents
      tuber_clod

Logo Categorization

logo-cluster


Introduction

From the highest number images folder, the logo ids and corresponding design ids are extracted from Logotool logos and designs collection using PyMongo Client. From the obtained design ids, information about the customer is extracted. Details about logos, customers and orders are available .

 

Documentation

https://javrasoftware.sharepoint.com/:w:/r/sites/rdinaiml/_layouts/15/doc.aspx?sourcedoc=%7Bf25118eb-339a-463d-a265-e2033ef870de%7D&file=image%20clustering%20script%20using%20clip%20and%20hierarchical%20clu.docx&action=default&mobileredirect=true

POC(only http )
http://poc.javra.com/

GIT

https://git.javra.com/internal-rnd/image-clustering-using-clip.git

Smart Weight App​

smartweight-APP

Introduction

The Smart Weight Serial Keyboard Application for Android is developed natively using Core Java, making it a powerful and efficient utility. The app acts as a bridge between a connected serial device and any target Android application, allowing users to capture data from the connected serial device and emulate keyboard input, thus ensuring compatibility with various Android applications that accept keyboard input. 

  • Bridge between a connected serial device and any android device
  • Automated Data Entry from weighing machine to Android 
  • Adding Bluetooth (in progress)

Overview
https://javra.com/Smart-Weight-Serial-Keyboard-Application

Download
https://play.google.com/store/apps/details?id=javra.WeightSmart

Documentation:-
https://javrasoftware.sharepoint.com/:w:/r/sites/RDinAIML/_layouts/15/Doc2.aspx?action=edit&sourcedoc=%7B4759eaa1-250f-4abf-a0c0-47d6274e82fd%7D&wdOrigin=TEAMS-MAGLEV.teamsSdk_ns.rwc&wdExp=TEAMS-TREATMENT&wdhostclicktime=1722935761190&web=1

GIT

https://git.javra.com/rnd/read-ohaus-data/-/tree/main?ref_type=heads