Spring 2023
Shape Shifting
Tools
Photography
P5JS
Adobe InDesign
Video Editing
3D Modeling
Description
A photographic exploration that focuses on the human silhouette. Through a series of experiments and activities, the project results in the development of a programmed script to detect and display the presence of a human silhouette against my pre-created camouflages. However, a playful twist arises as I attempt to deceive the very machine I created by introducing my created patterns. This resulted in a fun interplay between perception and deception.
Outcome
A script that captures the human body’s keypoints at varied confidences, in conjunction with pre-created camoflagues I created to truly understand how this machine learning model works.
Process
A set of activities preceded the final scripting and experiment with projection mapping and camoflague creation. They include the participation of others, their hands, their thoughts, their sharing of self.
Part 01
In the first experiment, I curated a collection of view windows featuring renowned artifacts, such as Stonehenge, the Nazca Lines, and the Sphinx of Giza, to name a few. These remarkable objects not only hold cultural significance but also possess an intriguing quality of ambiguity. Blending elements of fiction and reality, they challenge our perception and remain recognizable despite their enigmatic nature. Overlaying these views unveils a captivating revelation, showcasing varying degrees of clarity and inviting us to explore the delicate balance between recognition and the mysterious allure of the unknown.‘
Part 02
In the second experiment, I delve into the realm of individual isolation, capturing the solitary silhouette of a person. This exploration takes a twist as I embark on the intriguing endeavor of manipulating the image, with a particular focus on creating a camouflage pattern unique to that specific individual. Delving into the interplay between individual and shape.
Part 03
In experiment three, I wanted to know about people's perception of silhouette, and outline. I ask people to draw the prompt "draw an outline of your inner self" this prompt was chosen in order to understand something that feels un revealed and intrinsic to how people understand silhouette, but in a more metaphorical way in comparison to asking others to draw their inner selves, and that can be translated from 2D to 3D afterwards. I used collection boxes, so that participants may submit their drawings through the slot, and later played with the forms in 3D.
Part 04
The inverse of experiment three, asking participants the same prompt “an outline of your inner self” but this time in 3D Clay, to be trasnlated into digital space. I find it interesting to view such an object as an artifact. As the artifacts that we know today do not famously depict the human in great accuracy, but rather the public consciousness of the human being at that point of the time. There are a remarkable amount of shapes in which the human body has already been depicted. It was interesting to note people’s remarks about why they shaped clay the way they did, citing the way in which others perceive them. It is an irony, perhaps, that our idea of our innermost selves stems from ideas we collect often purely externally from others.
Part 05
Now turning from others, to my own self: What does an ‘analog’ double exposure read as? How can I use a self-projection to trick the eye?
Part 06
In the final experiment, I delved into the development of a script that utilized pre-trained machine learning models, implemented through ML5.js, to detect and indicate the presence of human silhouettes. ML5.js, a JavaScript library, simplifies the usage of pre-trained machine learning models in web applications. It acts as a bridge, enabling developers to perform tasks like image classification, object detection, and pose estimation without the need for extensive training. By assigning credibility scores to individual keypoints, the program provided detailed feedback on its perception. These credibility scores relied on a threshold value, a decimal number that determined when recognized keypoints were displayed and confirmed. Initially intending to create a digital garment concealing the silhouette from human perception using shape filters, I shifted gears and decided to test the program's capabilities. Through the final experiment, I explored animations, graphics, and sound-triggered elements to manipulate the program and assess its limitations. By adjusting the threshold and experimenting with motion, scale, division, color, and contrast, I gained valuable insights into the program's recognition capabilities within an artistic context, shedding light on the interplay between visual triggers and keypoints' confidence levels.
Side By Side:
Captioned with Explanations: