Joni Juvonen

Computer vision and machine learning software engineer.

[AMD Silo AI @AMDSiloAI]

Two of our AI experts, Mikko Tukiainen & Joni Juvonen @QitVision share their learnings from a @kaggle challenge aiming to improve prostate cancer diagnostics with computer vision: https://t.co/WM9JWc76Px #computervision #CV #ML #cancer #diagnostics

[Joni Juvonen @QitVision]

We're currently in the top 5 📈 #kaggle https://t.co/GPlqwlbWvJ

[AMD Silo AI @AMDSiloAI]

Our AI Engineer Joni Juvonen & AI Scientist Mikko Tukiainen are speaking about deepfakes & how to detect them with computer vision at https://t.co/BkubfJoxV9's upcoming meetup next Wednesday – secure your spot https://t.co/8YJKGKh65h #computervision #turku #CV #deepfake @turkuai

[AMD Silo AI @AMDSiloAI]

Our AI Engineer Joni Juvonen @QitVision & AI Scientist Mikko Tukiainen can’t get enough of exploring @kaggle's computer vision competitions. Read how the two AI experts use the community to learn about the latest #CV technologies. #keeplearning #AI https://t.co/XdO0es10nj

[AMD Silo AI @AMDSiloAI]

One of the key aspects in developing self-driving technology is the vehicle's perception: what the car "sees and hears". Our AI Engineer Joni Juvonen presents his #LIDAR U-net model for object detection he successfully built for an online challenge: https://t.co/TQkTioEscS #CV

[Joni Juvonen @QitVision]

H&E histopathological staining appearance augmentation for digital pathology. This is adapted from H&E normalization of Schaugf and Mitkovetta. https://t.co/sUc8ASc3HE
#digitalpathology #fastai

[Auria Biopankki @AuriaBiopankki]

@AuriaBiopankki'n data-analyytikkotiimi osallistui kilpailuun, jossa on tavoitteena kehittää #keinoäly luokitelemaan diabeteksen retinopatiaa vakavuusluokkiin silmänpohjakuvista. Tällä hetkellä tiimi on "hopeasijoilla" top 5%:ssä! #biopankki #kaggle - https://t.co/Dy1D6xWbi6

[Joni Juvonen @QitVision]

I won a Kaggle Swag prize with the top voted kernel in histopathologic cancer detection. Thanks to @kaggle, @fastdotai, and all who voted! I learned a great deal from that competition. https://t.co/67JtlLdz8U

[Joni Juvonen @QitVision]

@Etvi_J It works though...User tests ✅

[Joni Juvonen @QitVision]

Santa has an NP-hard optimization problem. https://t.co/pIpgygKJkm

[Joni Juvonen @QitVision]

@ReaktorNow Thanks for ruining my holidays 😅

[Joni Juvonen @QitVision]

Published a tutorial of an ML pipeline for detecting metastatic cancer. #fastai https://t.co/67JtlLdz8U

[Joni Juvonen @QitVision]

Ever wanted to add object detection to your videos? Here's a Python script for that. https://t.co/3PGCsgqPhH #TensorFlow #python

[Joni Juvonen @QitVision]

"the component waveform may be configured to the percent duty cycle and a processor for a predetermined in the art to provide a process that the substrate that is embodiment, and the controller in a processor element" 🤔

[Joni Juvonen @QitVision]

#AI learns to write patent descriptions 🙄:
"The dispersion in the control unit and processing interference is interrabled in which the the status of the module in discharged by the art compaters..."

#Keras #TensorFlow
https://t.co/HGOVWXIDVR

[Joni Juvonen @QitVision]

Combining neural style transfer and deep dream outputs. This one has a bit of Goncharova, Hokusai, and some deep dreams of MIT places. #styletransfer #deepdream

[Janelle Shane @JanelleCShane]

This program will try to draw pictures of whatever you describe and it's so delightful/terrifying.
https://t.co/AEarP8oPJU

[Joni Juvonen @QitVision]

3D printing AI-generated objects.

Source code and tutorial for 3D generation: https://t.co/T4zTYKUmW5
#3Dprinting #TensorFlow #AI

[Joni Juvonen @QitVision]

Generating 3D objects is surprisingly hard and unstable for generative adversarial nets (GAN), but with the recent relativistic cost functions, I managed to get stable training and quite nice chairs.

64-3D-RaSGAN code and tutorial with #tensorflow: https://t.co/T4zTYKUmW5

[Joni Juvonen @QitVision]

Juniper Video Library version 0.5 released. Faster and lighter neural network model for video content tagging and better previews.
https://t.co/jpRYMnuDzF #TensorFlow #opensource