LiquidHash aims to verify the authenticity of liquid food products without opening bottles and using only smartphone cameras. We leverage observed bubble characteristics -- size, shape, and speed -- to infer liquid authenticity.
We are witnessing a surge in the reported cases of counterfeit liquid products in the market including olive oil, honey, and alcohol. Counterfeiters often adulterate the liquid products by replacing a large portion of the authentic content with cheaper substitutes (e.g., mixing vodka with cheaper alcohol or potentially toxic methanol). Exacerbating the problem, the counterfeits are packaged and sealed to factory standards, rendering it extremely difficult for an average consumer to identify them. While solutions exist, they are often impractical for the general public as they require specialized and costly equipment. To overcome these limitations, we propose LiquidHash, a novel counterfeit liquid food product detection system. LiquidHash is a practical solution that only requires the use of a commodity smartphone to detect adulterated liquid products without opening the bottles. LiquidHash works by detecting and tracking the shape and movement of air bubbles that form inside the bottles. We implement LiquidHash and evaluate its feasibility with real-world experiments under varying conditions with a total of more than 500 minutes of video recording and observe an overall detection accuracy of up to 95%.
We envision that LiquidHash could be adopted by an average user with a simple bottle flipping.
(1) Start flipping
(2) Record bubbles using smartphone camera
(Optional) Achieve better performance and usability with a bottle cap accessory
(Optional) Proposed bottle cap accessory
Illustration of Selected Technical Challenges
To make LiquidHash work, we encounter several technical challenges. We illustrate two of them here.
(1) Noise in Measuring Bubble Characteristics: Inconsistent Bubble Trajectories
There are multiple sources of noise in measuring bubble characteristics, including rotation motion, camera placement, inconsistent bubble shapes, and inconsistent bubble trajectories. We illustrate inconsistent bubble trajectories here.
Erratic bubble trajectories arise before liquid stablizes, and uniform bubble trajectories after liquid stabilizes
(Non-desirable) Erratic and irregular bubble trajectories
(Desirable) Uniform and regular bubble trajectories
To tackle this challenge, we utilize image processing techniques to select and process input video frames to contain only steady frames, the frames containing bubbles moving in uniform and regular trajectories.
LiquidHash: Pre-processing module
(2) Minute Difference in Bubble Characteristics
The difference in bubble characteristics between authentic and adulterated liquid content could be extremely minute. In the following example, we can hardly see the difference between (1) extra virgin olive oil and (2) extra virgin olive oil adulterated with 30% sunflower oil.
The minute difference in observed bubble characteristics
(1) Extra Virgin Olive Oil
(2) Extra Virgin Olive Oil Adulterated with 30% Sunflower Oil
To tackle this challenge, we utilize computer vision techniques to segment bubble regions and obtain unique bubble trajectories across frames. Then we obtain bubble features representing bubble characteristics, namely size, shape, and speed, to optimize for machine learning classifiers.
LiquidHash: Bubble Feature Extraction Module