Looks as if the “Trek” game re-make is continuing … at least for a bit 😛

I got some coding time yesterday and so did some more work. Now the Long Range Scanner and Short Range Scanner work and you can navigate in either scanner view by simply tapping where you want to go … So progress 🙂

Also made things look a little bit prettier … though that might be my opinion and not yours 😛

Maybe next weekend I’ll get to do some more work, I guess we’ll see … While I think Flutter would have made this even more easy to code, SwiftUI has held its own so far …

#Game #StarTrek #Coding #SwiftUI #macOS
A screenshot from WIP “Trek” ga…

@troz I’ve made a bit more progress since I like working out the UI and figuring out how it should all work in this new version … But not sure if I’ll get around to completing it. Will let you know if I do 🙂
WIP screenshot from “Trek” game…

I think I’ve talked about the old “Trek” game on here at least once or twice 🙂

I think I might be obsessed with it — mostly because there is this one magical version I remember from the early 90s that I can’t seem to find anywhere. (Of course, the magic might be all in my mind … or in the fact that I can’t find it 😛)

I did some fiddling with a Mac OS 9 emulator this weekend and in the process ran across the “Trek” game again and it got me looking at “Trek” games in general again. In the process, I came across a version of the game for Windows 9x that is probably the most feature-filled version that I’ve seen (screenshot attached).

That made me think of re-writing a version of the game myself and I kind of got started last night. All I have so far is a plan for what I want it to look like … or at least half a plan (second screenshot).

Not sure if I’ll actually go through with completing it, but it’s a nice thing to dream about 🙂

#Game #StarTrek #Coding #SwiftUI
A screenshot from the game “Sta…
A screenshot from an early atte…

Decided to give DeepFloyd a try today on macOS.

The good news? It works … kinda 😛

The bad news? It doesn’t work all the way … as was to be expected 🙂

I took the following code from their GitHub repo (https://github.com/deep-floyd/IF) and modified for an Apple Silicon (M1) Mac. Here’s the actual code I ran:

from diffusers import DiffusionPipeline
from diffusers.utils import pt_to_pil
import torch

# stage 1
stage_1 = DiffusionPipeline.from_pretrained("DeepFloyd/IF-I-M-v1.0").to("mps")

# stage 2
stage_2 = DiffusionPipeline.from_pretrained("DeepFloyd/IF-II-M-v1.0", text_encoder=None).to("mps")

# stage 3
safety_modules = {"feature_extractor": stage_1.feature_extractor, "safety_checker": stage_1.safety_checker, "watermarker": stage_1.watermarker}
stage_3 = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-x4-upscaler", **safety_modules).to("mps")

prompt = 'a photo of a kangaroo wearing an orange hoodie and blue sunglasses standing in front of the eiffel tower holding a sign that says "very deep learning"'

# text embeds
prompt_embeds, negative_embeds = stage_1.encode_prompt(prompt)

generator = torch.manual_seed(0)

# stage 1
image = stage_1(prompt_embeds=prompt_embeds, negative_prompt_embeds=negative_embeds, generator=generator, output_type="pt").images
pt_to_pil(image)[0].save("./if_stage_I.png")

# stage 2
image = stage_2(image=image, prompt_embeds=prompt_embeds, negative_prompt_embeds=negative_embeds, generator=generator, output_type="pt").images
pt_to_pil(image)[0].save("./if_stage_II.png")

# stage 3
image = stage_3(prompt=prompt, image=image, generator=generator, noise_level=100).images
image[0].save("./if_stage_III.png")

You have to make sure that diffusers, transformers, and accelereate (at least in my own trial) are fully up-to-date. The larger models probably work too but it took too long to download/test and so I opted for the smallest models.

Stage I and II generated images but stage III errored out. I will need to figure out what happened there later …

Resulting images are attached …

#DeepLearning #MachineLearning #DeepFloyd #ImageGeneration

Stage I image — 64 x 64 in size…
Stage II image — 256 x 256 in s…

So … DOSBox …

On Friday, once work was done, I kind of started thinking about “SimLife”.

I’d posted about it a week or two ago and I thought it might be fun to try it out again and see if it was as bad as I remembered.

I went and installed DOSBox (https://www.dosbox.com/) but discovered that Boxer (http://boxerapp.com/) no longer worked for my version of macOS since it has not been updated in ages — shows how long it’s been since I tried DOSBox I guess 😛

That sent me down all sorts of byways and alleyways trying out various alternatives to Boxer and even setting up MS-DOS 6.22 on a VM but I eventually came back to DOSBox since it was the easiest/simplest way to get going.

I tried out “SimLife” again and was pleasantly surprised to find that I actually kind of liked the game. I probably hadn’t tried out the tutorial back in the day and probably just tried clicking around and it’s the kind of game where you really need to know the mechanics. So I might try playing it again over the next few weeks …

But that opening graphic, so gorgeous even for that time period — just so lovely 🙂

#Games #DOS #Memories #Sim #History
“SimLife” opening graphic / spl…
“SimLife” game start showing th…

Does anybody remember the classic “Star Trek” game from the 70’s? Or was that way before anybody’s time? 😛

I was playing around with DOSBox (more on that in a separate post) and was reminded of the game suddenly.

I recall playing a version on DOS back in 1990 (or 91) which had color graphics and more graphical controls than “Trek73” or “Super Trek”. The closest one I can find now is “EGA Trek” (see first screenshot) but I’m fairly certain that that wasn’t the one I played because the Main Viewer area in “EGA Trek” cycles through different graphics and I don’t remember that.

Also the game I played had long range and short range scanners which used up energy and that doesn’t seem to be a feature of “EGA Trek”.

Now that I’ve remembered the game, I really want to play that exact version again 😛

I’ve played “EGA Trek”, “VTrek” (second screenshot) and gone through a bunch of newer variations that people created on GitHub. But none of them have the same feel as the game I remember. If anybody knows of the game I’m talking of, please let me know?

I think part of the trouble is that back then they were all called “Star Trek” and that was it. So many variations by so many people.

In fact, I thought the one I liked might have been created in Turbo Pascal (for no reason I could think of) and almost went down another rabbit hole setting up Turbo Pascal just to re-create the game for DOS 😛

Then I realized that if I were to re-create it that I might as well do it in Swift or Flutter so that I can have it avaialbe on more platforms. Will I do it? I don’t know … I guess it depends on how badly I want to play the version I remember again 🙂

#Games #StarTrek #DOS #Memories #History
“EGA Trek” game screen showing …
“VTrek” game screen showing a b…

"Inversion-Based Style Transfer with Diffusion Models. (arXiv:2211.13203v2 [cs.CV] UPDATED)" — Accurately transfer the style attributes such as semantics, material, object shape, brushstrokes and colors of an input image to the target image.

Paper: http://arxiv.org/abs/2211.13203
Code: https://github.com/zyxelsa/InST

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Style transfer results by using…

"Scaling up GANs for Text-to-Image Synthesis. (arXiv:2303.05511v1 [cs.CV])" — Adapting the GAN architecture for generating images via text prompts to work with large datasets, offering faster generation times and higher resolutions.

Paper: http://arxiv.org/abs/2303.05511

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Some sample images generated by…

"Mark My Words: Dangers of Watermarked Images in ImageNet. (arXiv:2303.05498v1 [cs.LG])" — A look at the potential risks of watermarked images present in
ImageNet and their impact on popular Deep Neural Networks (DNN) trained on this dataset.

Paper: http://arxiv.org/abs/2303.05498

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Multiple images with watermarks…

"3DGen: Triplane Latent Diffusion for Textured Mesh Generation. (arXiv:2303.05371v1 [cs.CV])" — Generating 3D meshes using diffusion models.

Paper: http://arxiv.org/abs/2303.05371

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Single category unconditional m…

"BaDLAD: A Large Multi-Domain Bengali Document Layout Analysis Dataset. (arXiv:2303.05325v1 [cs.CV])" — A dataset to help with Bengali document layout analysis, containing 33,695 human annotated document samples from six domains.

Paper: http://arxiv.org/abs/2303.05325
Code: https://github.com/anon-user-for-web/badlad

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
A sample of the different layou…

"Detecting Images Generated by Diffusers. (arXiv:2303.05275v1 [cs.CV])" — Detecting if given images were generated by diffusion models or not. Appears to be limited to detecting images only if images generated by a particular diffusion model were in the training set.

Paper: http://arxiv.org/abs/2303.05275

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
The two different network setup…

"O2RNet: Occluder-Occludee Relational Network for Robust Apple Detection in Clustered Orchard Environments. (arXiv:2303.04884v1 [cs.CV])" — Using deep learning to detect apples in orchards for harvesting.

Paper: http://arxiv.org/abs/2303.04884

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Six sample images from the coll…

"Interactive Cartoonization with Controllable Perceptual Factors. (arXiv:2212.09555v2 [cs.CV] UPDATED)" — A model capable of converting input images to cartoons where you have the ability to control the texture and color of the final image.

Paper: http://arxiv.org/abs/2212.09555

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Example of the background makin…

"Visual ChatGPT: Talking, Drawing and Editing with Visual Foundation Models. (arXiv:2303.04671v1 [cs.CV])" — An amalgamation of ChatGPT and image generation models such as Stable Diffusion to allow providing ChatGPT with not just text input but also images and to be able to provide visual questions or image editing instructions to ChatGPT.

Paper: http://arxiv.org/abs/2303.04671
No repo exists on GitHub matching the one linked in the paper.

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Overview of Visual ChatGPT. The…

"A Prompt Log Analysis of Text-to-Image Generation Systems. (arXiv:2303.04587v1 [cs.HC])" — Analyzing prompts used with diffusion models to get an idea of the informational needs of the users and to get an idea as to how to improve text-to-image generation systems.

Paper: http://arxiv.org/abs/2303.04587

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Two tables showing most frequen…

"Diffusion in the Dark: A Diffusion Model for Low-Light Text Recognition. (arXiv:2303.04291v1 [eess.IV])" — Recognizing text in images taken under low-light conditions while preserving high-frequency details in the image.

Paper: http://arxiv.org/abs/2303.04291

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Low-light image reconstruction …

"EscherNet 101. (arXiv:2303.04208v1 [cs.CV])" — Teaching an artificial neural network symmetry groups and their relations as found in M. C. Escher drawings and other similar artwork.

Paper: http://arxiv.org/abs/2303.04208

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
(A) Three drawings by M.C. Esch…

"Neural Style Transfer for Vector Graphics. (arXiv:2303.03405v1 [cs.CV])" — Transferring style from an input style image to an source vector image to create a target vector image that adheres to the style of the input style image.

Paper: http://arxiv.org/abs/2303.03405

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
We propose a novel neural style…

"DLT: Conditioned layout generation with Joint Discrete-Continuous Diffusion Layout Transformer. (arXiv:2303.03755v1 [cs.CV])" — Using diffusion models to create visual layouts as part of graphics design.

Paper: http://arxiv.org/abs/2303.03755

#AI #CV #NewPaper #DeepLearning #MachineLearning

<<Find this useful? Please boost so that others can benefit too 🙂>>
Overview of our Joint Discrete-…