Chasing Beauty… With a Twist

The cosmetics industry is enormous. No, not that big. Even bigger! For every lipstick, there are equally as many blushes, concealers, and contours. If you think AI algorithms can be implemented in the field of cosmetics, this article is for you!

Chasing Beauty… With a Twist
Written by TechnoLynx Published on 29 Aug 2024

Introduction

Even though the basic characteristics are the same in the human species (2 arms, 2 legs, etc.), there are certain characteristics that make us as unique as the stars in the sky. Each of us has his own genes, which are expressed to create the wonderful individuals we are. When unpacking a gift, what we are looking forward to is the contents of the box, not the paper wrap. The same applies to humans. True beauty comes from within, but, as with presents, nice wrapping really sets the mood.

Make me Stunning

Throwback

The beauty industry has become a multibillion-dollar market, with new products from major brands and celebrity lineups emerging every day. Cosmetics have been around for ages, in fact, since ancient times. Excavations have shown that ancient Egyptians used to grind minerals such as green malachite or black galena to create makeup or used red ochre to blush their cheeks. The Iliad and the Odyssey were the first sources that showed that the Greeks infused plants, flowers, spices, and fragrant woods like myrrh and oregano with oil to create fragrances (Cosmetics in the Ancient World - World History Encyclopedia). Of course, some of these preparations proved to be poisonous or just gross, so we have stopped using them, but the hunt for external beauty still goes on.

Figure 1 – Charcoal placed in a mortar to be used as eyeliner, shadow, or eyebrow shaping (Med, 2018).
Figure 1 – Charcoal placed in a mortar to be used as eyeliner, shadow, or eyebrow shaping (Med, 2018).

Wanna Go Shopping?

So, you decided to walk into the mall. There it is, shiny, magnificent! The cosmetics store! You hear its call, and you just need to get inside. You see rows after rows of primers, foundations, contours, blushes, lipsticks, moisturisers, shades, concealers, and it goes on and on! We know that deep down, you want to try them all, but we also know how busy you are. Here is where it gets really interesting: AI comes to aid!

If you read our article on cosmetology, you have seen how smart mirrors can analyse your characteristics to find the perfect hairstyle for you. Who said that the same can’t be done when choosing cosmetics? Companies are developing smart mirrors specifically for this purpose. These powerful tools can be the best ally for both pros who want something extra and beginners who just don’t know where to start. The device uses GPU accelerated Computer Vision (CV) to scan your face in real time. Through multiple registers of different areas, the mirror is able to identify your skin tone, eye colour and shape, hairstyle, facial characteristics and symmetry, and even skin condition to make suggestions of potential combinations that work well together or skin routine products. Of course, you won’t be convinced just because a machine says so. What if the machine could show you the result? Yes, it’s possible with an Augmented Reality (AR) try-on. Use this branch of Extended Reality (XR) to have a look at yourself wearing makeup without even putting it on!

Figure 2 – AI & AR software used to analyse skin condition (left) and skin tone (right) developed by the company ‘Perfect’ (Smart Makeup Mirror: The Complete Guide 2024).
Figure 2 – AI & AR software used to analyse skin condition (left) and skin tone (right) developed by the company ‘Perfect’ (Smart Makeup Mirror: The Complete Guide 2024).

Unfortunately, so far, there has been no collaboration between those under-development devices and cosmetics brands, but if such a thing happens, things will get even better. An established connection to a cloud server of a company will be able to not only make suggestions on colours and hues that will turn you irresistible, but it will make the device capable of giving you specific information on exactly the brand and the colour code of the product. And just when you think things can’t get better, why don’t we introduce the Internet of Things (IoT) to this? You have found the product instantaneously, yet you need to use your phone or laptop to order it? Nope! Just press a button on your smart mirror, and the product will be delivered to you. Now, maybe you don’t want to spend the extra money on a niche device. Understandable. This implementation can be done using the powerful camera your smartphone has! Now that we mentioned phones, with all these AI tools that are constantly being implemented, why not make the ‘best’ even better? Natural Language Processing (NLP) models have gone far and beyond. You have seen them in the different GPT assistants that you use to do your homework or write an email for you. This, however, is not the first time you have encountered them. Digital assistants have used NLPs for a while to extract information and form answers (you may want to recall a certain fruit phone at this point). Now, take this and use a speech-to-text-to-speech algorithm, and there you have it. Your smart mirror will finally be able to answer when you ask, ‘Mirror mirror on the wall, who is the fairest of them all?’ and even give you feedback on the selection of products that will make you even more dazzling!

Read more: AI Revolutionising Fashion & Beauty

Why me?

There is a dark side to the cosmetics industry that is worth discussing, and this is no other than animal cruelty. Sadly enough, major cosmetics companies continue to test their products on animals or, to comply with local regulations, perform these tests in other countries where this practice is allowed (These Beauty Brands Are Still Tested on Animals, 2015). The reason behind that is the safety of the consumers, as no company wants to get involved in litigations; nevertheless, it is not fair to perform such tests on beings that cannot express themselves. People for the Ethical Treatment of Animals (PETA) is the largest animal rights organisation in the world, with more than 9 million members and supporters worldwide.

Figure 3 – Milestones reached in the United Kingdom for animal rights on cosmetic tests from the introduction of the first EU provision in 1993 to the full ban in 2013 (Cosmetics).
Figure 3 – Milestones reached in the United Kingdom for animal rights on cosmetic tests from the introduction of the first EU provision in 1993 to the full ban in 2013 (Cosmetics).

If you are wondering whether there are ways to bypass such practices for the well-being of our four-legged friends, there is, and technology is on our side! As you know, our skin has specific characteristics, some of which are common among us, while some of us have our own peculiarities. For example, most people have what we call ‘normal skin type’, but others have ‘oily’, ‘dry’ or even a combination of the above. We already mentioned how a well-trained CV algorithm can be used to process your skin and extract useful information to analyse it. However, this is applicable not only to humans. The same algorithm (with some tweaks here and there) can be applied for the analysis of animal skin. In this way, we could predict the skin reaction before it even occurs, completely excluding the animals from the process.

But wait, if you can analyse a skin to avoid the involvement of animals, why not analyse the human skin in the first place? Well, you are absolutely right! Remember when we said earlier that each one of us is unique? Through careful planning, an enormous database representing each skin could be created and constantly enriched with fresh data. Then, millions of virtual tests could be run in local infrastructure by Edge Computing to be able to support the huge volume of information, reaching conclusions with precision that can be identical to testing on animals. This does not necessarily mean that animals are out of the equation, thankfully, for a good purpose. As humans have peculiar skins, so do animals. A similar algorithm and tests could be run to develop cosmetics and skin products for our little friends. NLPs could also be involved in such processes. You have your skin analysis, and you are using your smartphone or smart mirror to choose the perfect cosmetics, yet there are so many options that you need to apply some filters when searching. This can be done instantly through NLPs. Just say which ingredients you absolutely want or don’t want the products to contain, and there you have it! Perfect choices every time!

It’s Just a Scar

Medicine is probably one of the most respected professions out there. There is a specific branch, though, that is specifically destined to make our looks better, and that is no other than cosmetic plastic surgery. Now, don’t be fooled. Plastic surgery is not just about improving our appearance for cosmetic purposes. There are cases where the surgery has a purely functional purpose. Examples of such cases include blepharoplasty, where the eyelid interferes with vision, rhinoplasty, where the nose is optimally shaped not only to perfectly match the rest of the face but to allow proper oxygenation of the body, or even scar revision, where surgeons can minimise the visibility of scars and improve their texture, giving a more natural appearance to the skin, potentially even improving mobility and relieving discomfort.

For sure, fashion and current trends certainly have an impact on what decisions we make about all sorts of matter, including plastic surgery. For example, an 86% increase in buttock lifting has been observed in women from 2019 to 2022, according to the American Society of Plastic Surgery, while breast reduction for aesthetic purposes has seen an increase of 54%, and blepharoplasty is at 13% in the same era (Plastic Surgery Statistics).

Figure 4 – A woman with lines drawn by the plastic surgeon as indicators of the sections that need to be made in surgery (Admin, 2024).
Figure 4 – A woman with lines drawn by the plastic surgeon as indicators of the sections that need to be made in surgery (Admin, 2024).

Does it Really Matter?

It doesn’t matter at all why someone chooses to have an operation of this type. It’s your body and, therefore, your choices. However, not all surgeries have 100% success. Sometimes, even routine operations can go south, with the implications being severe. We need to be sure about our doctor, as the doctor needs to be sure of what they are doing. Isn’t that, however, a bit unfair for the less experienced? By implementing AI in the process, success is guaranteed.

The usual suspect, CV, can get to work even before the operation starts. By carefully examining the patient, the doctor can accurately decide the optimal method, the exact dosage of the filling material, or the amount of tissue that must be removed, eliminating potential errors. Or inside the operation theatre, it wouldn’t hurt to have an extra pair of computer eyes watching over the procedure or even indicating to the surgeon exactly where they need to operate through AR or Virtual Reality (VR). Most likely, you are familiar with the concept of robotic surgery, with medical devices such as the world-famous DaVinci appearing on the news every now and then. These devices are mostly designed for larger-scale operations, where much detail and precision beyond human limits are needed. However, why limit things and exclude cosmetic surgery? A medical device that can perform cosmetic surgeries would not only be beneficial but also profitable. If we enrich it with AI technology to such an extent that it can almost perform the surgery on its own, the benefits will be even greater. Maximum precision, minimum scars and rehabilitation time, optimal results, and satisfied patients!

Read more: How Augmented Reality is Transforming Beauty and Cosmetics

Summing Up

The cosmetics industry can surely benefit from AI. Algorithms such as Computer Vision and Natural Language Processing can really transform the way we interact with cosmetics, both on a private and a corporate level. Using AI, we can have our own assistant when choosing cosmetics, eliminate animal cruelty, and achieve optimal cosmetic surgery results.

What We Offer

At TechnoLynx, we are all about innovation. We specialise in providing custom-tailored tech solutions specifically for your needs. We know the benefits of integrating AI into a wide variety of industries, including cosmetics, while ensuring safety in human-machine interactions, managing and analysing large data sets, and addressing ethical considerations.

Our precise software solutions are specifically designed to empower AI-driven algorithms in different fields and industries. At TechnoLynx, we are driven to adapt to the ever-changing AI landscape. Our solutions are specifically designed to increase efficiency, accuracy, and productivity for your business. Don’t hesitate, contact us. We will be more than happy to answer any questions!

List of references

Cost, Efficiency, and Value Are Not the Same Metric

Cost, Efficiency, and Value Are Not the Same Metric

17/04/2026

Performance per dollar. Tokens per watt. Cost per request. These sound like the same thing said differently, but they measure genuinely different dimensions of AI infrastructure economics. Conflating them leads to infrastructure decisions that optimize for the wrong objective.
Precision Is an Economic Lever in Inference Systems

Precision Is an Economic Lever in Inference Systems

17/04/2026

Precision isn't just a numerical setting — it's an economic one. Choosing FP8 over BF16, or INT8 over FP16, changes throughput, latency, memory footprint, and power draw simultaneously. For inference at scale, these changes compound into significant cost differences.
Precision Choices Are Constrained by Hardware Architecture

Precision Choices Are Constrained by Hardware Architecture

17/04/2026

You can't run FP8 inference on hardware that doesn't have FP8 tensor cores. Precision format decisions are conditional on the accelerator's architecture — its tensor core generation, native format support, and the efficiency penalties for unsupported formats.
Steady-State Performance, Cost, and Capacity Planning

Steady-State Performance, Cost, and Capacity Planning

17/04/2026

Capacity planning built on peak performance numbers over-provisions or under-delivers. Real infrastructure sizing requires steady-state throughput — the predictable, sustained output the system actually delivers over hours and days, not the number it hit in the first five minutes.
How Benchmark Context Gets Lost in Procurement

How Benchmark Context Gets Lost in Procurement

16/04/2026

A benchmark result starts with full context — workload, software stack, measurement conditions. By the time it reaches a procurement deck, all that context is gone. The failure mode is not wrong benchmarks but context loss during propagation.
Building an Audit Trail: Benchmarks as Evidence for Governance and Risk

Building an Audit Trail: Benchmarks as Evidence for Governance and Risk

16/04/2026

High-value AI hardware decisions need traceable evidence, not slide-deck bullet points. When benchmarks are documented with methodology, assumptions, and limitations, they become auditable institutional evidence — defensible under scrutiny and revisitable when conditions change.
The Comparability Protocol: Why Benchmark Methodology Defines What You Can Compare

The Comparability Protocol: Why Benchmark Methodology Defines What You Can Compare

16/04/2026

Two benchmark scores can only be compared if they share a declared methodology — the same workload, precision, measurement protocol, and reporting conditions. Without that contract, the comparison is arithmetic on numbers of unknown provenance.
A Decision Framework for Choosing AI Hardware

A Decision Framework for Choosing AI Hardware

16/04/2026

Hardware selection is a multivariate decision under uncertainty — not a score comparison. This framework walks through the steps: defining the decision, matching evaluation to deployment, measuring what predicts production, preserving tradeoffs, and building a repeatable process.
How Benchmarks Shape Organizations Before Anyone Reads the Score

How Benchmarks Shape Organizations Before Anyone Reads the Score

16/04/2026

Before a benchmark score informs a purchase, it has already shaped what gets optimized, what gets reported, and what the organization considers important. Benchmarks function as decision infrastructure — and that influence deserves more scrutiny than the number itself.
Accuracy Loss from Lower Precision Is Task‑Dependent

Accuracy Loss from Lower Precision Is Task‑Dependent

16/04/2026

Reduced precision does not produce a uniform accuracy penalty. Sensitivity depends on the task, the metric, and the evaluation setup — and accuracy impact cannot be assumed without measurement.
Precision Is a Design Parameter, Not a Quality Compromise

Precision Is a Design Parameter, Not a Quality Compromise

16/04/2026

Numerical precision is an explicit design parameter in AI systems, not a moral downgrade in quality. This article reframes precision as a representation choice with intentional trade-offs, not a concession made reluctantly.
Mixed Precision Works by Exploiting Numerical Tolerance

Mixed Precision Works by Exploiting Numerical Tolerance

16/04/2026

Not every multiplication deserves 32 bits. Mixed precision works because neural network computations have uneven numerical sensitivity — some operations tolerate aggressive precision reduction, others don't — and the performance gains come from telling them apart.

Throughput vs Latency: Choosing the Wrong Optimization Target

16/04/2026

Throughput and latency are different objectives that often compete for the same resources. This article explains the trade-off, why batch size reshapes behavior, and why percentiles matter more than averages in latency-sensitive systems.

Quantization Is Controlled Approximation, Not Model Damage

16/04/2026

When someone says 'quantize the model,' the instinct is to hear 'degrade the model.' That framing is wrong. Quantization is controlled numerical approximation — a deliberate engineering trade-off with bounded, measurable error characteristics — not an act of destruction.

GPU Utilization Is Not Performance

15/04/2026

The utilization percentage in nvidia-smi reports kernel scheduling activity, not efficiency or throughput. This article explains the metric's exact definition, why it routinely misleads in both directions, and what to pair it with for accurate performance reads.

FP8, FP16, and BF16 Represent Different Operating Regimes

15/04/2026

FP8 is not just 'half of FP16.' Each numerical format encodes a different set of assumptions about range, precision, and risk tolerance. Choosing between them means choosing operating regimes — different trade-offs between throughput, numerical stability, and what the hardware can actually accelerate.

Peak Performance vs Steady‑State Performance in AI

15/04/2026

AI systems rarely operate at peak. This article defines the peak vs. steady-state distinction, explains when each regime applies, and shows why evaluations that capture only peak conditions mischaracterize real-world throughput.

The Software Stack Is a First‑Class Performance Component

15/04/2026

Drivers, runtimes, frameworks, and libraries define the execution path that determines GPU throughput. This article traces how each software layer introduces real performance ceilings and why version-level detail must be explicit in any credible comparison.

The Mythology of 100% GPU Utilization

15/04/2026

Is 100% GPU utilization bad? Will it damage the hardware? Should you be worried? For datacenter AI workloads, sustained high utilization is normal — and the anxiety around it usually reflects gaming-era intuitions that don't apply.

Why Benchmarks Fail to Match Real AI Workloads

15/04/2026

The word 'realistic' gets attached to benchmarks freely, but real AI workloads have properties that synthetic benchmarks structurally omit: variable request patterns, queuing dynamics, mixed operations, and workload shapes that change the hardware's operating regime.

Why Identical GPUs Often Perform Differently

15/04/2026

'Same GPU' does not imply the same performance. This article explains why system configuration, software versions, and execution context routinely outweigh nominal hardware identity.

Training and Inference Are Fundamentally Different Workloads

15/04/2026

A GPU that excels at training may disappoint at inference, and vice versa. Training and inference stress different system components, follow different scaling rules, and demand different optimization strategies. Treating them as interchangeable is a design error.

Performance Ownership Spans Hardware and Software Teams

15/04/2026

When an AI workload underperforms, attribution is the first casualty. Hardware blames software. Software blames hardware. The actual problem lives in the gap between them — and no single team owns that gap.

Performance Emerges from the Hardware × Software Stack

15/04/2026

AI performance is an emergent property of hardware, software, and workload operating together. This article explains why outcomes cannot be attributed to hardware alone and why the stack is the true unit of performance.

Power, Thermals, and the Hidden Governors of Performance

14/04/2026

Every GPU has a physical ceiling that sits below its theoretical peak. Power limits, thermal throttling, and transient boost clocks mean that the performance you read on the spec sheet is not the performance the hardware sustains. The physics always wins.

Why AI Performance Changes Over Time

14/04/2026

That impressive throughput number from the first five minutes of a training run? It probably won't hold. AI workload performance shifts over time due to warmup effects, thermal dynamics, scheduling changes, and memory pressure. Understanding why is the first step toward trustworthy measurement.

CUDA, Frameworks, and Ecosystem Lock-In

14/04/2026

Why is it so hard to switch away from CUDA? Because the lock-in isn't in the API — it's in the ecosystem. Libraries, tooling, community knowledge, and years of optimization create switching costs that no hardware swap alone can overcome.

GPUs Are Part of a Larger System

14/04/2026

CPU overhead, memory bandwidth, PCIe topology, and host-side scheduling routinely limit what a GPU can deliver — even when the accelerator itself has headroom. This article maps the non-GPU bottlenecks that determine real AI throughput.

Why AI Performance Must Be Measured Under Representative Workloads

14/04/2026

Spec sheets, leaderboards, and vendor numbers cannot substitute for empirical measurement under your own workload and stack. Defensible performance conclusions require representative execution — not estimates, not extrapolations.

Low GPU Utilization: Where the Real Bottlenecks Hide

14/04/2026

When GPU utilization drops below expectations, the cause usually isn't the GPU itself. This article traces common bottleneck patterns — host-side stalls, memory-bandwidth limits, pipeline bubbles — that create the illusion of idle hardware.

Why GPU Performance Is Not a Single Number

14/04/2026

AI GPU performance is multi-dimensional and workload-dependent. This article explains why scalar rankings collapse incompatible objectives and why 'best GPU' questions are structurally underspecified.

What a GPU Benchmark Actually Measures

14/04/2026

A benchmark result is not a hardware measurement — it is an execution measurement. The GPU, the software stack, and the workload all contribute to the number. Reading it correctly requires knowing which parts of the system shaped the outcome.

Why Spec‑Sheet Benchmarking Fails for AI

14/04/2026

GPU spec sheets describe theoretical limits. This article explains why real AI performance is an execution property shaped by workload, software, and sustained system behavior.

Cracking the Mystery of AI’s Black Box

4/02/2026

A guide to the AI black box problem, why it matters, how it affects real-world systems, and what organisations can do to manage it.

Inside Augmented Reality: A 2026 Guide

3/02/2026

A 2026 guide explaining how augmented reality works, how AR systems blend digital elements with the real world, and how users interact with digital content through modern AR technology.

Smarter Checks for AI Detection Accuracy

2/02/2026

A clear guide to AI detectors, why they matter, how they relate to generative AI and modern writing, and how TechnoLynx supports responsible and high‑quality content practices.

Choosing Vulkan, OpenCL, SYCL or CUDA for GPU Compute

28/01/2026

A practical comparison of Vulkan, OpenCL, SYCL and CUDA, covering portability, performance, tooling, and how to pick the right path for GPU compute across different hardware vendors.

Deep Learning Models for Accurate Object Size Classification

27/01/2026

A clear and practical guide to deep learning models for object size classification, covering feature extraction, model architectures, detection pipelines, and real‑world considerations.

TPU vs GPU: Which Is Better for Deep Learning?

26/01/2026

A practical comparison of TPUs and GPUs for deep learning workloads, covering performance, architecture, cost, scalability, and real‑world training and inference considerations.

CUDA vs ROCm: Choosing for Modern AI

20/01/2026

A practical comparison of CUDA vs ROCm for GPU compute in modern AI, covering performance, developer experience, software stack maturity, cost savings, and data‑centre deployment.

Best Practices for Training Deep Learning Models

19/01/2026

A clear and practical guide to the best practices for training deep learning models, covering data preparation, architecture choices, optimisation, and strategies to prevent overfitting.

Measuring GPU Benchmarks for AI

15/01/2026

A practical guide to GPU benchmarks for AI; what to measure, how to run fair tests, and how to turn results into decisions for real‑world projects.

GPU‑Accelerated Computing for Modern Data Science

14/01/2026

Learn how GPU‑accelerated computing boosts data science workflows, improves training speed, and supports real‑time AI applications with high‑performance parallel processing.

CUDA vs OpenCL: Picking the Right GPU Path

13/01/2026

A clear, practical guide to cuda vs opencl for GPU programming, covering portability, performance, tooling, ecosystem fit, and how to choose for your team and workload.

Performance Engineering for Scalable Deep Learning Systems

12/01/2026

Learn how performance engineering optimises deep learning frameworks for large-scale distributed AI workloads using advanced compute architectures and state-of-the-art techniques.

Choosing TPUs or GPUs for Modern AI Workloads

10/01/2026

A clear, practical guide to TPU vs GPU for training and inference, covering architecture, energy efficiency, cost, and deployment at large scale across on‑prem and Google Cloud.

GPU vs TPU vs CPU: Performance and Efficiency Explained

10/01/2026

Understand GPU vs TPU vs CPU for accelerating machine learning workloads—covering architecture, energy efficiency, and performance for large-scale neural networks.

Energy-Efficient GPU for Machine Learning

9/01/2026

Learn how energy-efficient GPUs optimise AI workloads, reduce power consumption, and deliver cost-effective performance for training and inference in deep learning models.
Back See Blogs
arrow icon