unreal engine | 3D modeling | e-commerce
|Research Phase
A tool that provides a virtual model of the user to try on garments before making a purchase, aiming to deliver realistic cloth simulations that boost buyer confidence and reduce returns.
Current Stage: Research for available tools and basic implementation in UE5 to confirm feasibility
The tool should serve two purposes:
1. Allow users to conveniently try on clothes through a photorealistic avatar of themselves with realistic multilayer cloth simulation in real time.
2. Enable designers of all levels and backgrounds to publish their work, allowing users to try on and vote on designs even before a piece of clothing goes into production.
This concept is not entirely new in e-commerce; for example, Apple provides AR visualization of products on its website, and there have been multiple mobile app attempts at virtual try-ons. However, most of these solutions use limited hardware (for instance, the GPU power of a mobile device) and consequently fall short in both graphical fidelity and physical accuracy.
When clothing arrives and does not meet expectations, time and resources are wasted on returns.
Many existing virtual try-on solutions suffer from poor graphics and inaccurate simulations, making the simulated experience unrealistic. They remain at a somewhat gimmicky stage. Recent advancements in affordable, powerful GPUs—such as the GeForce RTX 5070 Ti—have made high-fidelity, real-time cloth simulation increasingly attainable for average consumers. This trend will continue and as these powerful graphics cards become cheaper and more accessible, the rise of advanced simulation tools for online shopping is on the horizon.
The technology can see widespread adoption for online shopping, potentially integrating with established e-commerce platforms and significantly improving the customer experience. As hardware barriers continue to drop, more users will benefit from accurate visualization and physics, resulting in fewer returns and higher customer satisfaction.
UI/UX design | prototyping | NeuroMuse
|Early Development
A mock-up design for NeuroMuse that repurposes current features and UI into an independent project, enhancing user experience in an existing or upcoming platform.
Current Stage: Buggy basic feature implementation
The near-term steps involve stabilizing features and refining the UI for clarity and ease of use. This project serves as a testing ground for interface concepts that may later be integrated into the main NeuroMuse platform.
automation | data processing | AI integration
|Near Completion
A workflow that streamlines repeated search tasks by automating repetitive queries and processes, designed to save time and improve consistency in data gathering.
Current Stage: Close to completion; no noticeable bugs; basic UI
This workflow streamlines repeated search tasks by automating repetitive queries and processes.
Efforts should focus on fine-tuning query methods, expanding scraper capabilities, and simplifying the UI for speed and usability.
game development | rhythm mechanics | unreal engine
|Concept Phase
Planned improvements to the existing Rhythm Game concept, focusing on enhancing the rhythm-based fighting mechanic with new features and gameplay elements.
Current Stage: Idea
Current Rhythm Game concept exists for a rhythm-based fighting mechanic. Planned updates include enemy creation, projectile design, UI, point system, and map movement challenges.
game development | storytelling | level design
|Concept Phase
Enhancements to the Stone Game concept, focusing on storytelling and robust enemy creation, with plans to develop core mechanics and a cohesive narrative.
Current Stage: Idea
Stone is a game concept focused on storytelling and robust enemy creation. The next steps involve developing core mechanics and a cohesive narrative.
game concept | symbiotic mechanics | character design
|Concept Phase
A game concept set in a distant future where humans, robots, and hyperactive plants coexist, featuring a protagonist who replaces body parts with symbiotic plants that grant new abilities.
Current Stage: Idea (inspired by the Stone Cage prototype)
Flower Bed is set in a distant future where humans, robots, and hyperactive plants coexist. The protagonist loses body parts and replaces them with symbiotic plants that confer new abilities—granting more power and versatility but gradually diminishing the protagonist's human aspects. Gameplay balances robotic speed and agility with plant-based strength and power, reminiscent of the duality between Venom and Eddie Brock (from Marvel) or the Parasyte concept in Japanese anime.
A major gameplay hook is that these symbiotic plants may behave differently based on whether they attach to specific robot parts, how much mass they've replaced, and the particular plant species involved. Plants might provide rope-like mechanics for springing to higher ledges, extra power in combat, or specialized grappling abilities.
In this distant future, humans have colonized a rocky planet using bioengineered seeds and drones to create a more livable environment. Over time, these plants evolved in unexpected ways, developing heightened aggression and complex behaviors. Conflict arises when new drones and robots are deployed to curtail plant overgrowth, pitting mechanized forces against rampant vegetation.
Symbiotic plant modifications grant various gameplay enhancements, with different plant varieties offering unique powers. Advanced movement or grappling features could provide an extra layer of strategy.
Inspired by Sifu's multiple-run structure, Flower Bed allows players to choose or create their own difficulty through builds choice, encouraging repeated playthroughs. The aesthetic aims to appeal to a wide audience by mixing smooth animations with stylized visuals.
AI tools | animation | game development
|Proof of Concept
A technique using ComfyUI and other AI tools to simplify the generation of unique animations, addressing inconsistencies observed in publicly available AI animation solutions.
Current Stage: Proof of concept, simple setup to confirm the feasibility
The approach uses ComfyUI (or similar tools) to simplify the generation of unique animations, addressing inconsistencies observed in publicly available AI animation solutions. Free/public AI animation tools such as Neuroframe or InVideo AI produce inconsistent frames, making them less suitable for cohesive game development.
The Flower Bed project demands a large number of unique animations and effects for both the main character and for enemies or backgrounds. This demand is likely common within game development process.
AI-assisted animation generation will help ensure style consistency across diverse abilities. The plan is:
By feeding structured key frames into ComfyUI, developers can rapidly produce animations and prototypes without painstaking manual drawing. This is beneficial for: